1_MC1 Introduction_NN1 :_: linguistic_JJ forms_NN2 and_CC functions_VVZ The_AT functions_NN2 of_IO language_NN1 The_AT analysis_NN1 of_IO discourse_NN1 is_VBZ ,_, necessarily_RR ,_, the_AT analysis_NN1 of_IO language_NN1 in_II use_NN1 ._. 
As_II such_DA ,_, it_PPH1 can_VM not_XX be_VBI restricted_VVN to_II the_AT description_NN1 of_IO linguistic_JJ forms_NN2 independent_JJ of_IO the_AT purposes_NN2 or_CC functions_NN2 which_DDQ those_DD2 forms_NN2 are_VBR designed_VVN to_TO serve_VVI in_II human_JJ affairs_NN2 ._. 
While_CS some_DD linguists_NN2 may_VM concentrate_VVI on_II determining_VVG the_AT formal_JJ properties_NN2 of_IO a_AT1 language_NN1 ,_, the_AT discourse_NN1 analyst_NN1 is_VBZ committed_VVN to_II an_AT1 investigation_NN1 of_IO what_DDQ that_DD1 language_NN1 is_VBZ used_VVN for_IF ._. 
While_CS the_AT formal_JJ approach_NN1 has_VHZ a_AT1 long_JJ tradition_NN1 ,_, manifested_VVN in_II innumerable_JJ volumes_NN2 of_IO grammar_NN1 ,_, the_AT functional_JJ approach_NN1 is_VBZ less_RGR well_RR documented_VVN ._. 
Attempts_NN2 to_TO provide_VVI even_RR a_AT1 general_JJ set_NN1 of_IO labels_NN2 for_IF the_AT principal_JJ functions_NN2 of_IO language_NN1 have_VH0 resulted_VVN in_II vague_JJ ,_, and_CC often_RR confusing_JJ ,_, terminology_NN1 ._. 
We_PPIS2 will_VM adopt_VVI only_RR two_MC terms_NN2 to_TO describe_VVI the_AT major_JJ functions_NN2 of_IO language_NN1 and_CC emphasise_VV0 that_CST this_DD1 division_NN1 is_VBZ an_AT1 analytic_JJ convenience_NN1 ._. 
It_PPH1 would_VM be_VBI unlikely_JJ that_CST ,_, on_II any_DD occasion_NN1 ,_, a_AT1 natural_JJ language_NN1 utterance_NN1 would_VM be_VBI used_VVN to_TO fulfil_VVI only_RR one_MC1 function_NN1 ,_, to_II the_AT total_JJ exclusion_NN1 of_IO the_AT other_JJ ._. 
That_DD1 function_NN1 which_DDQ language_NN1 serves_VVZ in_II the_AT expression_NN1 of_IO '_GE content_NN1 '_GE we_PPIS2 will_VM describe_VVI as_CSA transactional_JJ ,_, and_CC that_DD1 function_NN1 involved_JJ in_II expressing_VVG social_JJ relations_NN2 and_CC personal_JJ attitudes_NN2 we_PPIS2 will_VM describe_VVI as_CSA interactional_JJ ._. 
Our_APPGE distinction_NN1 ,_, '_GE transactional_JJ /_FO interactional_JJ '_GE ,_, stands_VVZ in_II general_JJ correspondence_NN1 to_II the_AT functional_JJ dichotomies_NN2 '_GE representative_NN1 /_FO expressive_JJ '_GE ,_, found_VVN in_II Bhler_NP1 (_( 1934_MC )_) ,_, '_GE referential_JJ /_FO emotive_JJ '_GE (_( Jakobson_NP1 ,_, 1960_MC )_) ,_, '_GE ideational_JJ /_FO interpersonal_JJ '_GE (_( Halliday_NP1 ,_, 1970b_FO )_) and_CC '_GE descriptive_JJ /_FO socialexpressive_JJ '_GE (_( Lyons_NP1 ,_, 1977_MC )_) ,_, The_AT transactional_JJ view_NN1 Linguists_NN2 and_CC linguistic_JJ philosophers_NN2 tend_VV0 to_TO adopt_VVI a_AT1 limited_JJ approach_NN1 to_II the_AT functions_NN2 of_IO language_NN1 in_II society_NN1 ._. 
While_CS they_PPHS2 frequently_RR acknowledge_VV0 that_DD1 language_NN1 may_VM be_VBI used_VVN to_TO perform_VVI many_DA2 communicative_JJ functions_NN2 ,_, they_PPHS2 nonetheless_RR make_VV0 the_AT general_JJ assumption_NN1 that_CST the_AT most_RGT important_JJ function_NN1 is_VBZ the_AT communication_NN1 of_IO information_NN1 ._. 
Thus_RR Lyons_NP1 (_( 1977_MC :_: 32_MC )_) observes_VVZ that_CST the_AT notion_NN1 of_IO communication_NN1 is_VBZ readily_RR used_JJ '_GE of_IO feelings_NN2 ,_, moods_NN2 and_CC attitudes_NN2 '_GE but_CCB suggests_VVZ that_CST he_PPHS1 will_VM be_VBI primarily_RR interested_JJ in_II '_GE the_AT intentional_JJ transmission_NN1 of_IO factual_JJ ,_, or_CC propositional_JJ ,_, information_NN1 '_GE ._. 
Similarly_RR Bennett_NP1 (_( 1976_MC :_: 5_MC )_) remarks_NN2 '_GE it_PPH1 seems_VVZ likely_JJ that_CST communication_NN1 is_VBZ primarily_RR a_AT1 matter_NN1 of_IO a_AT1 speaker_NN1 's_GE seeking_VVG either_RR to_TO inform_VVI a_AT1 hearer_NN1 of_IO something_PN1 or_CC to_TO enjoin_VVI some_DD action_NN1 upon_II him_PPHO1 '_GE ._. 
The_AT value_NN1 of_IO the_AT use_NN1 of_IO language_NN1 to_TO transmit_VVI information_NN1 is_VBZ well_RR embedded_VVN in_II our_APPGE cultural_JJ mythology_NN1 ._. 
We_PPIS2 all_DB believe_VV0 that_CST it_PPH1 is_VBZ the_AT faculty_NN1 of_IO language_NN1 which_DDQ has_VHZ enabled_VVN the_AT human_JJ race_NN1 to_TO develop_VVI diverse_JJ cultures_NN2 ,_, each_DD1 with_IW its_APPGE distinctive_JJ social_JJ customs_NN2 ,_, religious_JJ observances_NN2 ,_, laws_NN2 ,_, oral_JJ traditions_NN2 ,_, patterns_NN2 of_IO trading_NN1 ,_, and_RR31 so_RR32 on_RR33 ._. 
We_PPIS2 all_DB believe_VV0 ,_, moreover_RR ,_, that_CST it_PPH1 is_VBZ the_AT acquisition_NN1 of_IO written_JJ language_NN1 which_DDQ has_VHZ permitted_VVN the_AT development_NN1 within_II some_DD of_IO these_DD2 cultures_NN2 of_IO philosophy_NN1 ,_, science_NN1 and_CC literature_NN1 (_( see_VV0 Goody_NP1 ,_, 1977_MC )_) ._. 
We_PPIS2 all_DB believe_VV0 that_CST this_DD1 development_NN1 is_VBZ made_VVN possible_JJ by_II the_AT ability_NN1 to_TO transfer_VVI information_NN1 through_II the_AT use_NN1 of_IO language_NN1 ,_, which_DDQ enables_VVZ man_NN1 to_TO utilise_VVI the_AT knowledge_NN1 of_IO his_APPGE forebears_NN2 ,_, and_CC the_AT knowledge_NN1 of_IO other_JJ men_NN2 in_II other_JJ cultures_NN2 ._. 
We_PPIS2 shall_VM call_VVI the_AT language_NN1 which_DDQ is_VBZ used_VVN to_TO convey_VVI '_GE factual_JJ or_CC propositional_JJ information_NN1 '_GE primarily_RR transactional_JJ language_NN1 ._. 
In_II primarily_RR transactional_JJ language_NN1 we_PPIS2 assume_VV0 that_CST what_DDQ the_AT speaker_NN1 (_( or_CC writer_NN1 )_) has_VHZ primarily_RR in_II mind_NN1 is_VBZ the_AT efficient_JJ transference_NN1 of_IO information_NN1 ._. 
Language_NN1 used_VVN in_II such_DA a_AT1 situation_NN1 is_VBZ primarily_RR '_GE message_NN1 oriented_JJ '_GE ._. 
It_PPH1 is_VBZ important_JJ that_CST the_AT recipient_NN1 gets_VVZ the_AT informative_JJ detail_NN1 correct_JJ ._. 
Thus_RR if_CS a_AT1 policeman_NN1 gives_VVZ directions_NN2 to_II a_AT1 traveller_NN1 ,_, a_AT1 doctor_NN1 tells_VVZ a_AT1 nurse_NN1 how_RRQ to_TO administer_VVI medicine_NN1 to_II a_AT1 patient_NN1 ,_, a_AT1 householder_NN1 puts_VVZ in_II an_AT1 insurance_NN1 claim_NN1 ,_, a_AT1 shop_NN1 assistant_NN1 explains_VVZ the_AT relative_JJ merits_NN2 of_IO two_MC types_NN2 of_IO knitting_NN1 wool_NN1 ,_, or_CC a_AT1 scientist_NN1 describes_VVZ an_AT1 experiment_NN1 ,_, in_II each_DD1 case_NN1 it_PPH1 matters_VVZ that_CST the_AT speaker_NN1 should_VM make_VVI what_DDQ he_PPHS1 says_VVZ (_( or_CC writes_VVZ )_) clear_JJ ._. 
There_EX will_VM be_VBI unfortunate_JJ (_( even_RR disastrous_JJ )_) consequences_NN2 in_II the_AT real_JJ world_NN1 if_CS the_AT message_NN1 is_VBZ not_XX properly_RR understood_VVN by_II the_AT recipient_NN1 ._. 
The_AT interactional_JJ view_NN1 Whereas_CS linguists_NN2 ,_, philosophers_NN2 of_IO language_NN1 and_CC psycho-linguists_NN2 have_VH0 ,_, in_RR21 general_RR22 ,_, paid_VVD particular_JJ attention_NN1 to_II the_AT use_NN1 of_IO language_NN1 for_IF the_AT transmission_NN1 of_IO '_GE factual_JJ or_CC propositional_JJ information_NN1 '_GE ,_, sociologists_NN2 and_CC sociolinguists_NN2 have_VH0 been_VBN particularly_RR concerned_JJ with_IW the_AT use_NN1 of_IO language_NN1 to_TO establish_VVI and_CC maintain_VVI social_JJ relationships_NN2 ._. 
In_II sociological_JJ and_CC anthropological_JJ literature_NN1 the_AT phatic_JJ use_NN1 of_IO language_NN1 has_VHZ been_VBN frequently_RR commented_VVN on_II particularly_RR the_AT conventional_JJ use_NN1 of_IO language_NN1 to_TO open_VVI talk-exchanges_NN2 and_CC to_TO close_VVI them_PPHO2 ._. 
Conversational_JJ analysts_NN2 have_VH0 been_VBN particularly_RR concerned_JJ with_IW the_AT use_NN1 of_IO language_NN1 to_TO negotiate_VVI role-relationships_NN2 ,_, peer-solidarity_NN1 ,_, the_AT exchange_NN1 of_IO turns_NN2 in_II a_AT1 conversation_NN1 ,_, the_AT saving_NN1 of_IO face_NN1 of_IO both_RR speaker_NN1 and_CC hearer_NN1 (_( cf._VV0 Labov_NP1 ,_, 1972a_FO ;_; Brown_NP1 and_CC Levinson_NP1 ,_, 1978_MC ;_; Sacks_NP1 ,_, Schegloff_NP1 &amp;_CC Jefferson_NP1 ,_, 1974_MC ;_; Lakoff_NP1 ,_, 1973_MC )_) ._. 
It_PPH1 is_VBZ clearly_RR the_AT case_NN1 that_CST a_AT1 great_JJ deal_NN1 of_IO everyday_JJ human_JJ interaction_NN1 is_VBZ characterised_VVN by_II the_AT primarily_RR interpersonal_JJ rather_II21 than_II22 the_AT primarily_RR transactional_JJ use_NN1 of_IO language_NN1 ._. 
When_CS two_MC strangers_NN2 are_VBR standing_VVG shivering_VVG at_II a_AT1 bus-stop_NN1 in_II an_AT1 icy_JJ wind_NN1 and_CC one_PN1 turns_VVZ to_II the_AT other_JJ and_CC says_58 '_GE My_APPGE goodness_NN1 ,_, it_PPH1 's_VBZ cold_JJ '_GE ,_, it_PPH1 is_VBZ difficult_JJ to_TO suppose_VVI that_CST the_AT primary_JJ intention_NN1 of_IO the_AT speaker_NN1 is_VBZ to_TO convey_VVI information_NN1 ._. 
It_PPH1 seems_VVZ much_DA1 more_DAR reasonable_JJ to_TO suggest_VVI that_CST the_AT speaker_NN1 is_VBZ indicating_VVG a_AT1 readiness_NN1 to_TO be_VBI friendly_JJ and_CC to_TO talk_VVI ._. 
Indeed_RR a_AT1 great_JJ deal_NN1 of_IO ordinary_JJ everyday_JJ conversation_NN1 appears_VVZ to_TO consist_VVI of_IO one_MC1 individual_NN1 commenting_VVG on_II something_PN1 which_DDQ is_VBZ present_JJ to_TO both_RR him_PPHO1 and_CC his_APPGE listener_NN1 ._. 
The_AT weather_NN1 is_VBZ of_RR21 course_RR22 the_AT most_RGT quoted_JJ example_NN1 of_IO this_DD1 in_II British_JJ English_NN1 ._. 
However_RRQV a_AT1 great_JJ deal_NN1 of_IO casual_JJ conversation_NN1 contains_VVZ phrases_NN2 and_CC echoes_NN2 of_IO phrases_NN2 which_DDQ appear_VV0 more_RRR to_TO be_VBI intended_VVN as_CSA contributions_NN2 to_II a_AT1 conversation_NN1 than_CSN to_TO be_VBI taken_VVN as_CSA instances_NN2 of_IO information-giving_NN1 ._. 
Thus_RR a_AT1 woman_NN1 on_II a_AT1 bus_NN1 describing_VVG the_AT way_NN1 a_AT1 mutual_JJ friend_NN1 has_VHZ been_VBN behaving_VVG ,_, getting_VVG out_II21 of_II22 bed_NN1 too_RG soon_RR after_II an_AT1 operation_NN1 ,_, concludes_VVZ her_APPGE turn_NN1 in_II the_AT conversation_NN1 by_II saying_VVG Aye_UH ,_, she_PPHS1 's_VBZ an_AT1 awfy_JJ woman._NNU (_( awfy_NN1 =_FO Sc_FO awful_JJ )_) This_DD1 might_VM be_VBI taken_VVN as_II an_AT1 informative_JJ summary_NN1 ._. 
Her_APPGE neighbour_NN1 then_RT says_VVZ reflectively_RR (_( having_VHG been_VBN supportively_RR uttering_VVG aye_UH ,_, aye_UH throughout_II the_AT first_MD speaker_NN1 's_GE turn_NN1 )_) :_: Aye_UH ,_, she_PPHS1 's_VBZ an_AT1 awfy_JJ woman_NN1 ._. 
Pirsig_NP1 (_( 1976_MC :_: 313_MC )_) remarks_NN2 of_IO such_DA a_AT1 conversation_NN1 :_: '_GE the_AT conversation_NN1 's_GE pace_NN1 intrigues_NN2 me_PPIO1 ._. 
It_PPH1 is_VBZ n't_XX intended_VVN to_TO go_VVI anywhere_RL ,_, just_RR fill_VV0 the_AT time_NNT1 of_IO day_NNT1 ..._... on_RP and_CC on_RP and_CC on_RP with_IW no_AT point_NN1 or_CC purpose_NN1 other_II21 than_II22 to_TO fill_VVI the_AT time_NNT1 ,_, like_II the_AT rocking_NN1 of_IO a_AT1 chair_NN1 ._. 
'_" What_DDQ seems_VVZ to_TO be_VBI primarily_RR at_II issue_NN1 here_RL is_VBZ the_AT sharing_NN1 of_IO a_AT1 common_JJ point_NN1 of_IO view_NN1 ._. 
Brown_NP1 &amp;_CC Levinson_NP1 point_VV0 out_RP the_AT importance_NN1 for_IF social_JJ relationships_NN2 of_IO establishing_VVG common_JJ ground_NN1 and_CC agreeing_VVG on_II points_NN2 of_IO view_NN1 ,_, and_CC illustrate_VV0 the_AT lengths_NN2 to_II which_DDQ speakers_NN2 in_II different_JJ cultures_NN2 will_VM go_VVI to_TO maintain_VVI an_AT1 appearance_NN1 of_IO agreement_NN1 ,_, and_CC they_PPHS2 remark_NN1 '_GE agreement_NN1 may_VM also_RR be_VBI stressed_VVN by_II repeating_JJ part_NN1 or_CC all_DB of_IO what_DDQ the_AT preceding_JJ speaker_NN1 has_VHZ said_VVN '_GE (_( 1978:117_MC )_) Whereas_CS ,_, as_CSA we_PPIS2 shall_VM note_VVI ,_, written_JJ language_NN1 is_VBZ ,_, in_RR21 general_RR22 ,_, used_VVN for_IF primarily_RR transactional_JJ purposes_NN2 ,_, it_PPH1 is_VBZ possible_JJ to_TO find_VVI written_JJ genres_NN2 whose_DDQGE purpose_NN1 is_VBZ not_XX primarily_RR to_TO inform_VVI but_CCB to_TO maintain_VVI social_JJ relationships-'thank_NN1 you'_NNU letters_NN2 ,_, love_VV0 letters_NN2 ,_, games_NN2 of_IO consequences_NN2 ,_, etc_RA ._. 
Spoken_VVN and_CC written_JJ language_NN1 Manner_NN1 of_IO production_NN1 From_II the_AT point_NN1 of_IO view_NN1 of_IO production_NN1 ,_, it_PPH1 is_VBZ clear_JJ that_CST spoken_JJ and_CC written_JJ language_NN1 make_VV0 somewhat_RR different_JJ demands_NN2 on_II language-producers_NN2 ._. 
The_AT speaker_NN1 has_VHZ available_JJ to_II him_PPHO1 the_AT full_JJ range_NN1 of_IO '_GE voice_NN1 quality_NN1 '_GE effects_NN2 (_( as_II31 well_II32 as_II33 facial_JJ expression_NN1 ,_, postural_JJ and_CC gestural_JJ systems_NN2 )_) ._. 
Armed_VVN with_IW these_DD2 he_PPHS1 can_VM always_RR override_VVI the_AT effect_NN1 of_IO the_AT words_NN2 he_PPHS1 speaks_VVZ ._. 
Thus_RR the_AT speaker_NN1 who_PNQS says_58 '_GE I_PPIS1 'd_VM really_RR like_VVI to_TO '_GE ,_, leaning_VVG forward_RL ,_, smiling_VVG ,_, with_IW a_AT1 '_GE warm_JJ ,_, breathy_JJ '_GE voice_NN1 quality_NN1 ,_, is_VBZ much_RR more_RGR likely_JJ to_TO be_VBI interpreted_VVN as_CSA meaning_VVG what_DDQ he_PPHS1 says_VVZ ,_, than_CSN another_DD1 speaker_NN1 uttering_VVG the_AT same_DA words_NN2 ,_, leaning_VVG away_RL ,_, brow_NN1 puckered_VVD ,_, with_IW a_AT1 '_GE sneering_NN1 ,_, nasal_NN1 '_GE voice_NN1 quality_NN1 ._. 
These_DD2 paralinguistic_JJ cues_NN2 are_VBR denied_VVN to_II the_AT writer_NN1 ._. 
We_PPIS2 shall_VM generally_RR ignore_VVI paralinguistic_JJ features_NN2 in_II spoken_JJ language_NN1 in_II this_DD1 book_NN1 since_CS the_AT data_NN we_PPIS2 shall_VM quote_VVI from_II is_VBZ spoken_VVN by_II co-operative_JJ adults_NN2 who_PNQS are_VBR not_XX exploiting_VVG paralinguistic_JJ resources_NN2 against_II the_AT verbal_JJ meanings_NN2 of_IO their_APPGE utterances_NN2 but_CCB are_VBR ,_, rather_RR ,_, using_VVG them_PPHO2 to_TO reinforce_VVI the_AT meaning_NN1 ._. 
Not_XX only_RR is_VBZ the_AT speaker_NN1 controlling_VVG the_AT production_NN1 of_IO communicative_JJ systems_NN2 which_DDQ are_VBR different_JJ from_II those_DD2 controlled_VVN by_II the_AT writer_NN1 ,_, he_PPHS1 is_VBZ also_RR processing_VVG that_DD1 production_NN1 under_II circumstances_NN2 which_DDQ are_VBR considerably_RR more_RGR demanding_JJ ._. 
The_AT speaker_NN1 must_VM monitor_VVI what_DDQ it_PPH1 is_VBZ that_CST he_PPHS1 has_VHZ just_RR said_VVN ,_, and_CC determine_VV0 whether_CSW it_PPH1 matches_VVZ his_APPGE intentions_NN2 ,_, while_CS he_PPHS1 is_VBZ uttering_VVG his_APPGE current_JJ phrase_NN1 and_CC monitoring_VVG that_CST ,_, and_CC simultaneously_RR planning_VVG his_APPGE next_MD utterance_NN1 and_CC fitting_VVG that_CST into_II the_AT overall_JJ pattern_NN1 of_IO what_DDQ he_PPHS1 wants_VVZ to_TO say_VVI and_CC monitoring_NN1 ,_, moreover_RR ,_, not_XX only_RR his_APPGE own_DA performance_NN1 but_CCB its_APPGE reception_NN1 by_II his_APPGE hearer_NN1 ._. 
He_PPHS1 has_VHZ no_AT permanent_JJ record_NN1 of_IO what_DDQ he_PPHS1 has_VHZ said_VVN earlier_RRR ,_, and_CC only_RR under_RG unusual_JJ circumstances_NN2 does_VDZ he_PPHS1 have_VHI notes_NN2 which_DDQ remind_VV0 him_PPHO1 what_DDQ he_PPHS1 wants_VVZ to_TO say_VVI next_MD ._. 
The_AT writer_NN1 ,_, on_II the_AT contrary_NN1 ,_, may_VM look_VVI over_II what_DDQ he_PPHS1 has_VHZ already_RR written_VVN ,_, pause_NN1 between_II each_DD1 word_NN1 with_IW no_AT fear_NN1 of_IO his_APPGE interlocutor_NN1 interrupting_VVG him_PPHO1 ,_, take_VV0 his_APPGE time_NNT1 in_II choosing_VVG a_AT1 particular_JJ word_NN1 ,_, even_RR looking_VVG it_PPH1 up_RP in_II the_AT dictionary_NN1 if_CS necessary_JJ ,_, check_VV0 his_APPGE progress_NN1 with_IW his_APPGE notes_NN2 ,_, reorder_VV0 what_DDQ he_PPHS1 has_VHZ written_VVN ,_, and_CC even_RR change_VV0 his_APPGE mind_NN1 about_II what_DDQ he_PPHS1 wants_VVZ to_TO say_VVI ._. 
Whereas_CS the_AT speaker_NN1 is_VBZ under_RG considerable_JJ pressure_NN1 to_TO keep_VVI on_RP talking_VVG during_II the_AT period_NN1 allotted_VVN to_II him_PPHO1 ,_, the_AT writer_NN1 is_VBZ characteristically_RR under_II no_AT such_DA pressure_NN1 ._. 
Whereas_CS the_AT speaker_NN1 knows_VVZ that_CST any_DD words_NN2 which_DDQ pass_VV0 his_APPGE lips_NN2 will_VM be_VBI heard_VVN by_II his_APPGE interlocutor_NN1 and_CC ,_, if_CS they_PPHS2 are_VBR not_XX what_DDQ he_PPHS1 intends_VVZ ,_, he_PPHS1 will_VM have_VHI to_TO undertake_VVI active_JJ ,_, public_NN1 '_GE repair_NN1 '_GE ,_, the_AT writer_NN1 can_VM cross_VVI out_RP and_CC rewrite_VVI in_II the_AT privacy_NN1 of_IO his_APPGE study_NN1 ._. 
There_EX are_VBR ,_, of_RR21 course_RR22 ,_, advantages_NN2 for_IF the_AT speaker_NN1 ._. 
He_PPHS1 can_VM observe_VVI his_APPGE interlocutor_NN1 and_CC ,_, if_CS he_PPHS1 wishes_VVZ to_TO ,_, modify_VV0 what_DDQ he_PPHS1 is_VBZ saying_VVG to_TO make_VVI it_PPH1 more_RGR accessible_JJ or_CC acceptable_JJ to_II his_APPGE hearer_NN1 ._. 
The_AT writer_NN1 has_VHZ no_AT access_NN1 to_II immediate_JJ feedback_NN1 and_CC simply_RR has_VHZ to_TO imagine_VVI the_AT reader_NN1 's_GE reaction_NN1 ._. 
It_PPH1 is_VBZ interesting_JJ to_TO observe_VVI the_AT behaviour_NN1 of_IO individuals_NN2 when_CS given_VVN a_AT1 choice_NN1 of_IO conducting_VVG a_AT1 piece_NN1 of_IO business_NN1 in_II person_NN1 or_CC in_II writing_NN1 ._. 
Under_II some_DD circumstances_NN2 a_AT1 face-to-face_JJ interaction_NN1 is_VBZ preferred_VVN but_CCB ,_, in_II others_NN2 ,_, for_IF a_AT1 variety_NN1 of_IO different_JJ reasons_NN2 ,_, the_AT individual_NN1 may_VM prefer_VVI to_TO conduct_VVI his_APPGE transaction_NN1 in_II writing_NN1 ._. 
Whereas_CS in_II a_AT1 spoken_JJ interaction_NN1 the_AT speaker_NN1 has_VHZ the_AT advantage_NN1 of_IO being_VBG able_JK to_TO monitor_VVI his_APPGE listener_NN1 's_GE minute-by-minute_JJ reaction_NN1 to_II what_DDQ he_PPHS1 says_VVZ ,_, he_PPHS1 also_RR suffers_VVZ from_II the_AT disadvantage_NN1 of_IO exposing_VVG his_APPGE own_DA feelings_NN2 (_( 'leaking'_VV0 ;_; Ekman_NP1 &amp;_CC Friesen_NP1 ,_, 1969_MC )_) and_CC of_IO having_VHG to_TO speak_VVI clearly_RR and_CC concisely_RR and_CC make_VVI immediate_JJ response_NN1 to_II whichever_DDQV way_NN1 his_APPGE interlocutor_NN1 reacts_VVZ ._. 
The_AT representation_NN1 of_IO discourse_NN1 :_: texts_NN2 So_RG far_RR we_PPIS2 have_VH0 considered_VVN in_II very_RG general_JJ terms_NN2 some_DD of_IO the_AT differences_NN2 in_II the_AT manner_NN1 of_IO production_NN1 of_IO writing_NN1 and_CC speech_NN1 ._. 
Before_CS we_PPIS2 go_VV0 on_RP to_TO discuss_VVI some_DD of_IO the_AT ways_NN2 in_II which_DDQ the_AT forms_NN2 of_IO speech_NN1 and_CC writing_NN1 differ_VV0 ,_, we_PPIS2 shall_VM consider_VVI ,_, in_II the_AT next_MD two_MC sections_NN2 ,_, some_DD of_IO the_AT problems_NN2 of_IO representing_VVG written_JJ and_CC spoken_JJ language_NN1 ._. 
We_PPIS2 shall_VM place_VVI this_DD1 within_II a_AT1 general_JJ discussion_NN1 of_IO what_DDQ it_PPH1 means_VVZ to_TO represent_VVI '_GE a_AT1 text_NN1 '_GE ._. 
We_PPIS2 shall_VM use_VVI text_NN1 as_II a_AT1 technical_JJ term_NN1 ,_, to_TO refer_VVI to_II the_AT verbal_JJ record_NN1 of_IO a_AT1 communicative_JJ act_NN1 ._. 
(_( For_IF another_DD1 approach_NN1 to_II text_NN1 cf._VV0 discussion_NN1 in_II Chapter_NN1 6_MC ._. )_) 
Written_JJ texts_NN2 The_AT notion_NN1 of_IO '_GE text_NN1 '_GE as_II a_AT1 printed_JJ record_NN1 is_VBZ familiar_JJ in_II the_AT study_NN1 of_IO literature_NN1 ._. 
A_ZZ1 '_GE text_NN1 '_GE may_VM be_VBI differently_RR presented_VVN in_II different_JJ editions_NN2 ,_, with_IW different_JJ type-face_NN1 ,_, on_II different_JJ sizes_NN2 of_IO paper_NN1 ,_, in_II one_MC1 or_CC two_MC columns_NN2 ,_, and_CC we_PPIS2 still_RR assume_VV0 ,_, from_II one_MC1 edition_NN1 to_II the_AT next_MD ,_, that_CST the_AT different_JJ presentations_NN2 all_DB represent_VV0 the_AT same_DA '_GE text_NN1 '_GE ._. 
It_PPH1 is_VBZ important_JJ to_TO consider_VVI just_RR what_DDQ it_PPH1 is_VBZ that_DD1 is_58 '_GE the_AT same_DA '_GE ._. 
Minimally_RR ,_, the_AT words_NN2 should_VM be_VBI the_AT same_DA words_NN2 ,_, presented_VVN in_II the_AT same_DA order_NN1 ._. 
Where_CS there_EX are_VBR disputed_JJ readings_NN2 of_IO texts_NN2 ,_, editors_NN2 usually_RR feel_VV0 obliged_VVN to_TO comment_VVI on_II the_AT crux_NN1 ;_; so_RR of_IO Hamlet_NP1 's_GE O_ZZ1 ,_, that_CST this_DD1 too_RG too_RG sullied_JJ flesh_NN1 would_VM melt_VVI (_( 1._MC ii.129_FO )_) Dover_NP1 Wilson_NP1 makes_VVZ it_PPH1 clear_JJ that_CST this_DD1 is_VBZ an_AT1 interpretation_NN1 ,_, since_CS the_AT second_MD Quarto_NN1 gives_58 '_GE too_RG too_RG sallied_JJ '_GE and_CC the_AT first_MD Folio_NN1 '_GE too_RG too_RG solid_JJ '_GE (_( Dover_NP1 Wilson_NP1 ,_, 1934_MC )_) ._. 
Even_RR where_CS there_EX is_VBZ no_AT doubt_NN1 about_II the_AT identity_NN1 of_IO words_NN2 and_CC their_APPGE correct_JJ sequence_NN1 ,_, replicating_VVG these_DD2 alone_RR does_VDZ not_XX guarantee_VVI an_AT1 adequate_JJ representation_NN1 of_IO a_AT1 text_NN1 ._. 
Consider_VV0 the_AT following_JJ extract_NN1 of_IO dialogue_NN1 from_II Pride_NP1 and_CC Prejudice_NN1 :_: '_GE Mr._NNB Bennet_NP1 ,_, how_RRQ can_VM you_PPY abuse_VVI your_APPGE own_DA children_NN2 in_II such_DA a_AT1 way_NN1 ?_? 
You_PPY take_VV0 delight_NN1 in_II vexing_VVG me_PPIO1 ._. 
You_PPY have_VH0 no_AT compassion_NN1 on_II my_APPGE poor_JJ nerves_NN2 ._. 
'_" '_" You_PPY mistake_VV0 me_PPIO1 ,_, my_APPGE dear_NN1 ._. 
I_PPIS1 have_VH0 a_AT1 high_JJ respect_NN1 for_IF your_APPGE nerves_NN2 ._. 
They_PPHS2 are_VBR my_APPGE old_JJ friends_NN2 ._. 
I_PPIS1 have_VH0 heard_VVN you_PPY mention_VVI them_PPHO2 with_IW consideration_NN1 these_DD2 twenty_MC years_NNT2 at_RR21 least_RR22 ._. 
'_" It_PPH1 is_VBZ clear_JJ that_CST more_DAR than_CSN simply_RR reproducing_VVG the_AT words_NN2 in_II their_APPGE correct_JJ order_NN1 is_VBZ required_VVN ._. 
It_PPH1 is_VBZ necessary_JJ to_TO replicate_VVI punctuation_NN1 conventions_NN2 ,_, as_II31 well_II32 as_II33 the_AT lineation_NN1 which_DDQ indicates_VVZ the_AT change_NN1 of_IO speaker_NN1 ._. 
The_AT extract_NN1 reads_VVZ as_CSA gobbledygook_NN1 if_CS it_PPH1 is_VBZ read_VVN as_II a_AT1 speech_NN1 by_II one_MC1 individual_NN1 ._. 
An_AT1 adequate_JJ representation_NN1 of_IO a_AT1 text_NN1 must_VM assign_VVI speeches_NN2 to_II the_AT correct_JJ characters_NN2 ,_, sentences_NN2 to_II the_AT correct_JJ paragraphs_NN2 ,_, and_CC paragraphs_NN2 to_II the_AT correct_JJ chapters_NN2 ._. 
The_AT author_NN1 's_GE organisation_NN1 and_CC staging_NN1 of_IO his_APPGE work_NN1 must_VM be_VBI preserved_VVN ._. 
In_II a_AT1 piece_NN1 of_IO expository_JJ prose_NN1 ,_, the_AT author_NN1 's_GE indication_NN1 of_IO the_AT development_NN1 of_IO the_AT argument_NN1 contributes_VVZ to_II the_AT reader_NN1 's_GE experience_NN1 of_IO the_AT text_NN1 ._. 
Thus_RR titles_NN2 ,_, chapter_NN1 headings_NN2 ,_, sub-divisions_NN2 and_CC sub-headings_NN2 all_DB indicate_VV0 to_II the_AT reader_NN1 how_RRQ the_AT author_NN1 intends_VVZ his_APPGE argument_NN1 to_TO be_VBI chunked_VVN ._. 
The_AT detail_NN1 of_IO lineation_NN1 rarely_RR matters_VVZ in_II expository_JJ or_CC descriptive_JJ prose_NN1 ._. 
However_RRQV it_PPH1 clearly_RR becomes_VVZ crucial_JJ in_II the_AT reproduction_NN1 of_IO poetry_NN1 ._. 
The_AT work_NN1 of_IO those_DD2 seventeenth-century_JJ poets_NN2 who_PNQS created_VVD poems_NN2 in_II the_AT shape_NN1 of_IO diamonds_NN2 or_CC butterflies_NN2 would_VM be_VBI largely_RR incomprehensible_JJ if_CS the_AT form_NN1 were_VBDR not_XX preserved_VVN ._. 
The_AT notion_NN1 of_IO '_GE text_NN1 '_GE reaches_NN2 beyond_II the_AT reproduction_NN1 of_IO printed_JJ material_NN1 in_II some_DD further_JJR printed_JJ form_NN1 ._. 
A_AT1 letter_NN1 ,_, handwritten_VVN in_II purple_JJ ink_NN1 with_IW many_DA2 curlicues_NN2 ,_, may_VM have_VHI its_APPGE text_NN1 reproduced_VVN in_II printed_JJ form_NN1 ._. 
Similarly_RR ,_, neutral_JJ printed_JJ versions_NN2 may_VM be_VBI produced_VVN of_IO handwritten_JJ shopping_NN1 lists_NN2 ,_, slogans_NN2 spray-painted_NN1 on_II21 to_II22 hoardings_NN2 ,_, and_CC public_JJ notices_NN2 embossed_VVN on_II metal_NN1 plates_NN2 ._. 
In_II each_DD1 case_NN1 the_AT '_GE text_NN1 '_GE will_NN1 be_VBI held_VVN to_TO have_VHI been_VBN reproduced_VVN if_CS the_AT words_NN2 ,_, the_AT punctuation_NN1 and_CC ,_, where_CS relevant_JJ ,_, the_AT lineation_NN1 are_VBR reproduced_VVN accurately_RR ._. 
Where_CS the_AT original_JJ text_NN1 exploits_NN2 typographical_JJ variety_NN1 ,_, a_AT1 text_NN1 reproduced_VVN in_II one_MC1 type-face_NN1 may_VM lack_VVI some_DD of_IO the_AT quality_NN1 of_IO the_AT original_JJ ._. 
An_AT1 obvious_JJ example_NN1 is_VBZ a_AT1 newspaper_NN1 item_NN1 which_DDQ may_VM exploit_VVI several_DA2 different_JJ type-faces_NN2 ,_, different_JJ sizes_NN2 of_IO type_NN1 and_CC a_AT1 particular_JJ shape_NN1 of_IO layout_NN1 ._. 
It_PPH1 is_VBZ interesting_JJ to_TO observe_VVI that_CST publishers_NN2 regularly_RR reproduce_VV0 conscious_JJ manipulation_NN1 of_IO the_AT written_JJ medium_NN1 on_II41 the_II42 part_II43 of_II44 the_AT writer_NN1 ._. 
Thus_RR Jane_NP1 Austen_NP1 's_GE expression_NN1 of_IO contrast_NN1 is_VBZ reproduced_VVN by_II publishers_NN2 in_II italics_NN2 :_: '_" Nay_UH ,_, '_GE said_VVD Elizabeth_NP1 ,_, '_" this_DD1 is_VBZ not_XX fair_JJ ._. 
You_PPY wish_VV0 to_TO think_VVI all_DB the_AT world_NN1 respectable_JJ ,_, and_CC are_VBR hurt_VVN if_CS I_PPIS1 speak_VV0 ill_RR of_IO any_DD body_NN1 ._. 
I_PPIS1 only_RR want_VV0 to_TO think_VVI you_PPY perfect_VV0 ..._... 
'_GE Similarly_RR Queen_NNB Victoria_NP1 's_GE use_NN1 of_IO underlining_VVG in_II her_APPGE handwritten_JJ journal_NN1 is_VBZ represented_VVN by_II her_APPGE publishers_NN2 in_II the_AT printed_JJ version_NN1 with_IW an_AT1 italic_JJ type-face_NN1 to_TO represent_VVI the_AT emphasis_NN1 she_PPHS1 wishes_VVZ to_TO indicate_VVI when_RRQ writing_NN1 of_IO Lord_NNB Melbourne_NP1 :_: he_PPHS1 gave_VVD me_PPIO1 such_DA a_AT1 kind_NN1 ,_, and_CC I_PPIS1 may_VM say_VVI ,_, fatherly_JJ look_NN1 (_( Thursday_NPD1 ,_, 28_MC June_NPM1 1838_MC )_) Where_RRQ the_AT writer_NN1 is_VBZ deliberately_RR exploiting_VVG the_AT resources_NN2 of_IO the_AT written_JJ medium_NN1 ,_, it_PPH1 seems_VVZ reasonable_JJ to_TO suggest_VVI that_CST that_DD1 manipulation_NN1 constitutes_VVZ part_NN1 of_IO the_AT text_NN1 ._. 
A_AT1 further_JJR illustration_NN1 of_IO this_DD1 is_VBZ to_TO be_VBI found_VVN in_II the_AT conventions_NN2 governing_VVG spelling_NN1 ._. 
In_RR21 general_RR22 we_PPIS2 assume_VV0 that_CST words_NN2 have_VH0 a_AT1 standardised_JJ spelling_NN1 in_II British_JJ English_NN1 ._. 
The_AT fact_NN1 of_IO the_AT standardisation_NN1 enables_VVZ authors_NN2 to_TO manipulate_VVI idiosyncratic_JJ spelling_NN1 to_TO achieve_VVI special_JJ effects_NN2 ._. 
Thus_RR in_II Winnie-the-Pooh_NP1 the_AT publishers_NN2 reproduce_VV0 the_AT notice_NN1 outside_II Owl_NN1 's_GE house_NN1 in_II one_MC1 inset_NN1 line_NN1 ,_, using_VVG capitals_NN2 ,_, and_CC with_IW the_AT author_NN1 's_GE own_DA spelling_NN1 :_: PLEZ_NP1 CNOKE_NP1 IF_CS AN_AT1 RNSR_NP1 IS_VBZ NOT_XX REQID_JJ The_AT point_NN1 that_CST the_AT author_NN1 makes_VVZ with_IW this_DD1 particular_JJ spelling_NN1 would_VM be_VBI lost_VVN if_CS the_AT words_NN2 were_VBDR reproduced_VVN in_II their_APPGE standard_JJ form_NN1 ._. 
It_PPH1 might_VM then_RT be_VBI claimed_VVN that_CST such_DA a_AT1 form_NN1 of_IO the_AT text_NN1 was_VBDZ incomplete_JJ or_CC inadequate_JJ ,_, because_CS the_AT point_NN1 which_DDQ the_AT author_NN1 wishes_VVZ to_TO make_VVI is_VBZ no_RR21 longer_RR22 accessible_JJ from_II the_AT written_JJ text_NN1 ._. 
Indeed_RR the_AT importance_NN1 of_IO the_AT correct_JJ citing_NN1 of_IO an_AT1 author_NN1 's_GE spelling_NN1 is_VBZ regularly_RR marked_VVN by_II the_AT insertion_NN1 of_IO sic_RR into_II a_AT1 citation_NN1 by_II a_AT1 second_MD author_NN1 who_PNQS wishes_VVZ to_TO disclaim_VVI responsibility_NN1 for_IF an_AT1 aberrant_JJ spelling_NN1 ._. 
We_PPIS2 have_VH0 so_RG far_RR been_VBN making_VVG the_AT simplifying_JJ assumption_NN1 that_CST it_PPH1 is_VBZ clear_JJ ,_, in_II all_DB cases_NN2 ,_, what_DDQ the_AT original_JJ text_NN1 consists_VVZ of_IO ._. 
Where_CS handwritten_JJ texts_NN2 are_VBR at_II issue_NN1 ,_, it_PPH1 is_VBZ often_RR the_AT case_NN1 that_CST the_AT individual_NN1 reproducing_VVG the_AT text_NN1 in_II a_AT1 printed_JJ version_NN1 has_VHZ to_TO make_VVI a_AT1 considerable_JJ effort_NN1 of_IO interpretation_NN1 to_TO assign_VVI a_AT1 value_NN1 to_II some_DD of_IO the_AT less_RGR legible_JJ words_NN2 ._. 
In_II literature_NN1 ,_, as_CSA we_PPIS2 have_VH0 remarked_VVN already_RR ,_, uncertainty_NN1 may_VM give_VVI rise_NN1 to_II cruces_NN2 ,_, to_II disputed_JJ texts_NN2 ._. 
In_II letters_NN2 ,_, prescriptions_NN2 ,_, shopping_NN1 lists_NN2 ,_, school_NN1 essays_NN2 ,_, the_AT reader_NN1 normally_RR pushes_VVZ through_II a_AT1 once-for-all_DB interpretation_NN1 of_IO a_AT1 text_NN1 which_DDQ may_VM never_RR be_VBI read_VVN again_RT ._. 
It_PPH1 must_VM be_VBI clear_JJ however_RR ,_, that_CST a_AT1 printed_JJ version_NN1 of_IO a_AT1 handwritten_JJ text_NN1 is_VBZ ,_, in_II an_AT1 important_JJ sense_NN1 ,_, an_AT1 interpretation_NN1 ._. 
This_DD1 is_VBZ particularly_RR clear_JJ in_II the_AT handwritten_JJ attempts_NN2 of_IO very_RG young_JJ children_NN2 where_RRQ the_AT adult_NN1 is_VBZ obliged_VVN to_TO assign_VVI each_DD1 large_JJ painstakingly_RR formed_VVN letter_NN1 token_NN1 to_II a_AT1 particular_JJ type_NN1 of_IO letter_NN1 ,_, which_DDQ he_PPHS1 may_VM then_RT re-interpret_VVI in_II41 the_II42 light_II43 of_II44 the_AT larger_JJR message_NN1 ._. 
Thus_RR we_PPIS2 have_VH0 before_II us_PPIO2 a_AT1 page_NN1 with_IW a_AT1 drawing_NN1 of_IO a_AT1 large_JJ animal_NN1 (_( reported_VVN to_TO be_VBI a_AT1 lion_NN1 )_) and_CC a_AT1 table_NN1 with_IW a_AT1 goldfish_NN bowl_NN1 on_II it_PPH1 ._. 
The_AT five-year-old_NN1 writes_VVZ below_II what_DDQ might_VM be_VBI transliterated_VVN as_II :_: 1_MC1 the_AT lion_NN1 wos_NN2 the_AT fish_NN to_II ti_NN2 it_PPH1 2_MC the_AT cat_NN1 wants_VVZ to_TO get_VVI dwon_NN1 the_AT steis_NN1 3_MC with_IW qwt_NNU to_II dsthhb_NNU thelion_NN1 A_ZZ1 possible_JJ interpretation_NN1 of_IO the_AT text_NN1 thus_RR represented_VVN might_VM be_VBI :_: The_AT lion_NN1 wants_VVZ the_AT fish_NN ,_, to_TO eat_VVI it_PPH1 ._. 
The_AT cat_NN1 wants_VVZ to_TO get_VVI down_II the_AT stairs_NN2 without_RL to_TO disturb_VVI the_AT lion_NN1 ._. 
The_AT transliteration_NN1 of_IO the_AT original_JJ with_IW qwt_NNU ,_, in_II line_NN1 3_MC ,_, reasonably_RR accurately_RR represents_VVZ the_AT first_MD letter_NN1 (_( which_DDQ might_VM also_RR be_VBI represented_VVN as_II a_AT1 figure_NN1 nine_MC if_CS nine_MC has_VHZ a_AT1 straight_JJ back_NN1 stroke_NN1 )_) ._. 
A_AT1 more_RGR charitable_JJ and_CC interpretive_JJ transliteration_NN1 would_VM render_VVI it_PPH1 as_II a_AT1 (_( i.e._REX '_GE unhatted_JJ '_GE a_AT1 with_IW a_AT1 long_JJ backstroke_NN1 (_( a_AT1 ._. 
)_) ._. We_PPIS2 shall_VM return_VVI to_II the_AT problem_NN1 of_IO the_AT interpretive_JJ work_NN1 of_IO the_AT reader_NN1 /_FO listener_NN1 in_II identifying_VVG the_AT words_NN2 which_DDQ constitute_VV0 the_AT text_NN1 ,_, in_II the_AT next_MD section_NN1 ._. 
Spoken_JJ texts_NN2 The_AT problems_NN2 encountered_VVN with_IW the_AT notion_NN1 of_IO '_GE text_NN1 '_GE as_CSA the_AT verbal_JJ record_NN1 of_IO a_AT1 communicative_JJ act_NN1 become_VV0 a_RR31 good_RR32 deal_RR33 more_RGR complex_JJ when_CS we_PPIS2 consider_VV0 what_DDQ is_VBZ meant_VVN by_II spoken_JJ '_GE text_NN1 '_GE ._. 
The_AT simplest_JJT view_NN1 to_TO assume_VVI is_VBZ that_CST a_AT1 tape-recording_NN1 of_IO a_AT1 communicative_JJ act_NN1 will_VM preserve_VVI the_AT '_GE text_NN1 '_GE ._. 
The_AT tape-recording_NN1 may_VM also_RR preserve_VVI a_RR31 good_RR32 deal_RR33 that_DD1 may_VM be_VBI extraneous_JJ to_II the_AT text_NN1 coughing_VVG ,_, chairs_NN2 creaking_JJ ,_, buses_NN2 going_VVG past_RL ,_, the_AT scratch_NN1 of_IO a_AT1 match_NN1 lighting_VVG a_AT1 cigarette_NN1 ._. 
We_PPIS2 shall_VM insist_VVI that_CST these_DD2 events_NN2 do_VD0 not_XX constitute_VVI part_NN1 of_IO the_AT text_NN1 (_( though_CS they_PPHS2 may_VM form_VVI part_NN1 of_IO the_AT relevant_JJ context_NN1 ,_, cf._VV0 Chapter_NN1 2_MC )_) ._. 
In_RR21 general_RR22 the_AT discourse_NN1 analyst_NN1 works_VVZ with_IW a_AT1 tape-recording_NN1 of_IO an_AT1 event_NN1 ,_, from_II which_DDQ he_PPHS1 then_RT makes_VVZ a_AT1 written_JJ transcription_NN1 ,_, annotated_VVN according_II21 to_II22 his_APPGE interests_NN2 on_II a_AT1 particular_JJ occasion_NN1 transcriptions_NN2 of_IO the_AT sort_NN1 which_DDQ will_VM appear_VVI in_II this_DD1 book_NN1 ._. 
He_PPHS1 has_VHZ to_TO determine_VVI what_DDQ constitutes_VVZ the_AT verbal_JJ event_NN1 ,_, and_CC what_DDQ form_VV0 he_PPHS1 will_VM transcribe_VVI it_PPH1 in_RP ._. 
Unless_CS the_AT analyst_NN1 produces_VVZ a_AT1 fine-grained_JJ phonetic_JJ transcription_NN1 (_( which_DDQ very_RG few_DA2 people_NN would_VM be_VBI able_JK to_TO read_VVI fluently_RR )_) details_NN2 of_IO accent_NN1 and_CC pronunciation_NN1 are_VBR lost_VVN ._. 
In_RR21 general_RR22 ,_, analysts_NN2 represent_VV0 speech_NN1 using_VVG normal_JJ orthographic_JJ conventions_NN2 ._. 
The_AT analyst_NN1 may_VM hear_VVI an_AT1 utterance_NN1 which_DDQ might_VM be_VBI transcribed_VVN phonemically_RR as_CSA /_FO greipbritn_VV0 /_FO ._. 
Is_VBZ he_PPHS1 to_TO render_VVI this_DD1 orthographically_RR as_CSA grape_NN1 britain_NN1 ?_? 
Hardly_RR ._. 
He_PPHS1 will_VM interpret_VVI what_DDQ he_PPHS1 hears_VVZ and_CC normalise_VVI to_II the_AT conventional_JJ orthographic_JJ form_NN1 Great_NP1 Britain_NP1 inserting_VVG conventional_JJ word_NN1 boundaries_NN2 in_II the_AT orthographic_JJ version_NN1 which_DDQ do_VD0 not_XX ,_, of_RR21 course_RR22 ,_, exist_VV0 in_II the_AT acoustic_JJ signal_NN1 ._. 
If_CS he_PPHS1 hears_VVZ a_AT1 form_NN1 /_FO gene_NN1 /_FO ,_, is_VBZ he_PPHS1 to_TO render_VVI this_DD1 in_II the_AT orthography_NN1 as_CSA gon_NN1 na_FW (_( which_DDQ for_IF some_DD readers_NN2 may_VM have_VHI a_AT1 peculiarly_RR American_JJ association_NN1 )_) or_CC gointuh_NN1 or_CC going_VVG to_II ?_? 
The_AT problem_NN1 is_VBZ a_AT1 very_RG real_JJ one_PN1 ,_, because_CS most_DAT speakers_NN2 constantly_RR simplify_VV0 words_NN2 phonetically_RR in_II the_AT stream_NN1 of_IO speech_NN1 (_( see_VV0 Brown_NP1 ,_, 1977_MC :_: ch._NNU 4_MC )_) ._. 
If_CS the_AT analyst_NN1 normalises_VVZ to_II the_AT conventional_JJ written_JJ form_NN1 ,_, the_AT words_NN2 take_VV0 on_RP a_AT1 formality_NN1 and_CC specificity_NN1 which_DDQ necessarily_RR misrepresent_JJ the_AT spoken_JJ form_NN1 ._. 
Problems_NN2 with_IW representing_VVG the_AT segmental_JJ record_NN1 of_IO the_AT words_NN2 spoken_VVN pale_JJ into_II insignificance_NN1 compared_VVN with_IW the_AT problems_NN2 of_IO representing_VVG the_AT suprasegmental_JJ record_NN1 (_( details_NN2 of_IO intonation_NN1 and_CC rhythm_NN1 )_) ._. 
We_PPIS2 have_VH0 no_AT standard_JJ conventions_NN2 for_IF representing_VVG the_AT paralinguistic_JJ features_NN2 of_IO the_AT utterance_NN1 which_DDQ are_VBR summarised_VVN as_58 '_GE voice_NN1 quality_NN1 '_GE ,_, yet_RR the_AT effect_NN1 of_IO an_AT1 utterance_NN1 being_VBG said_VVN kindly_RR and_CC sympathetically_RR is_VBZ clearly_RR very_RG different_JJ from_II the_AT effect_NN1 if_CS it_PPH1 is_VBZ said_VVN brutally_RR and_CC harshly_RR ._. 
Similarly_RR it_PPH1 is_VBZ usually_RR possible_JJ to_TO determine_VVI from_II a_AT1 speaker_NN1 's_GE voice_NN1 his_APPGE or_CC her_APPGE sex_NN1 ,_, approximate_JJ age_NN1 and_CC educational_JJ status_NN1 ,_, as_II31 well_II32 as_II33 some_DD aspects_NN2 of_IO state_NN1 of_IO health_NN1 and_CC personality_NN1 (_( see_VV0 Abercrombie_NP1 ,_, I968_FO ;_; Laver_NP1 ,_, 1980_MC )_) ._. 
It_PPH1 is_VBZ not_XX customary_JJ to_TO find_VVI any_DD detail_NN1 relating_VVG to_II these_DD2 indexical_JJ features_NN2 of_IO the_AT speaker_NN1 in_II transcriptions_NN2 by_II discourse_NN1 analysts_NN2 ._. 
In_RR21 general_RR22 ,_, too_RR ,_, rhythmic_JJ and_CC temporal_JJ features_NN2 of_IO speech_NN1 are_VBR ignored_VVN in_II transcriptions_NN2 ;_; the_AT rhythmic_JJ structure_NN1 which_DDQ appears_VVZ to_TO bind_VVI some_DD groups_NN2 of_IO words_NN2 more_RGR closely_RR together_RL than_CSN others_NN2 ,_, and_CC the_AT speeding_VVG up_RP and_CC slowing_VVG down_RP of_IO the_AT overall_JJ pace_NN1 of_IO speech_NN1 relative_II21 to_II22 the_AT speaker_NN1 's_GE normal_JJ pace_NN1 in_II a_AT1 given_JJ speech_NN1 situation_NN1 ,_, are_VBR such_DA complex_JJ variables_NN2 that_CST we_PPIS2 have_VH0 very_RG little_DA1 idea_NN1 how_RRQ they_PPHS2 are_VBR exploited_VVN in_II speech_NN1 and_CC to_II what_DDQ effect_NN1 (_( but_CCB ,_, cf._VV0 Butterworth_NP1 ,_, 1980_MC )_) ._. 
It_PPH1 seems_VVZ reasonable_JJ to_TO suggest_VVI ,_, though_CS ,_, that_CST these_DD2 variables_NN2 ,_, together_RL with_IW pause_NN1 and_CC intonation_NN1 ,_, perform_VV0 the_AT functions_NN2 in_II speech_NN1 that_CST punctuation_NN1 ,_, capitalization_NN1 ,_, italicization_NN1 ,_, paragraphing_NN1 etc._RA perform_VV0 in_II written_JJ language_NN1 ._. 
If_CS they_PPHS2 constitute_VV0 part_NN1 of_IO the_AT textual_JJ record_NN1 in_II written_JJ language_NN1 ,_, they_PPHS2 should_VM be_VBI included_VVN as_II part_NN1 of_IO the_AT textual_JJ record_NN1 in_II spoken_JJ language_NN1 ._. 
If_CS it_PPH1 is_VBZ relevant_JJ to_TO indicate_VVI Queen_NNB Victoria_NP1 's_VBZ underlining_VVG ,_, then_RT it_PPH1 is_VBZ surely_RR also_RR relevant_JJ to_TO indicate_VVI ,_, for_REX21 example_REX22 ,_, a_AT1 speaker_NN1 's_GE use_NN1 of_IO high_JJ pitch_NN1 and_CC loudness_NN1 to_TO indicate_VVI emphasis_NN1 ._. 
The_AT response_NN1 of_IO most_DAT analysts_NN2 to_II this_DD1 complex_JJ problem_NN1 is_VBZ to_TO present_VVI their_APPGE transcriptions_NN2 of_IO the_AT spoken_JJ text_NN1 using_VVG the_AT conventions_NN2 of_IO the_AT written_JJ language_NN1 ._. 
Thus_RR Cicourel_NP1 (_( 1973_MC )_) reproduces_VVZ three_MC utterances_NN2 recorded_VVN in_II a_AT1 classroom_NN1 in_II the_AT following_JJ way_NN1 :_: 1_MC1 Ci_FO :_: Like_II this_DD1 ?_? 2_MC T_ZZ1 :_: Okay_RR ,_, yeah_UH ,_, all_RR21 right_RR22 ,_, now_RT ..._... 3_MC Ri_NP1 :_: Now_RT what_DDQ are_VBR we_PPIS2 going_VVG to_TO do_VDI ?_? 
In_II I_ZZ1 and_CC 3_MC we_PPIS2 have_VH0 to_TO assume_VVI that_CST the_AT ?_? indicates_VVZ that_CST the_AT utterance_NN1 functions_NN2 as_II a_AT1 question_NN1 whether_CSW it_PPH1 is_VBZ formally_RR marked_VVN by_II ,_, for_REX21 instance_REX22 ,_, rising_VVG intonation_NN1 in_II the_AT case_NN1 of_IO I_ZZ1 ,_, we_PPIS2 are_VBR not_XX told_VVN ._. 
Similarly_RR the_AT status_NN1 of_IO commas_NN2 in_II the_AT speech_NN1 of_IO the_AT T(eacher)_NN1 is_VBZ not_XX made_VVN explicit_JJ presumably_RR they_PPHS2 are_VBR to_TO indicate_VVI pauses_NN2 in_II the_AT stream_NN1 of_IO speech_NN1 ,_, but_CCB it_PPH1 may_VM be_VBI that_CST they_PPHS2 simply_RR indicate_VV0 a_AT1 complex_NN1 of_IO rhythmic_JJ and_CC intonational_JJ cues_NN2 which_DDQ the_AT analyst_NN1 is_VBZ responding_VVG to_II ._. 
What_DDQ must_VM be_VBI clear_JJ in_II a_AT1 transcript_NN1 of_IO this_DD1 kind_NN1 is_VBZ that_CST a_AT1 great_JJ deal_NN1 of_IO interpretation_NN1 by_II the_AT analyst_NN1 has_VHZ gone_VVN on_RP before_II the_AT reader_NN1 encounters_VVZ this_58 '_GE data_NN '_GE ._. 
If_CS the_AT analyst_NN1 chooses_VVZ to_TO italicise_VVI a_AT1 word_NN1 in_II his_APPGE transcription_NN1 to_TO indicate_VVI ,_, for_REX21 example_REX22 ,_, the_AT speaker_NN1 's_GE high_JJ pitch_NN1 and_CC increased_JJ loudness_NN1 ,_, he_PPHS1 has_VHZ performed_VVN an_AT1 interpretation_NN1 on_II the_AT acoustic_JJ signal_NN1 ,_, an_AT1 interpretation_NN1 which_DDQ ,_, he_PPHS1 has_VHZ decided_VVN ,_, is_VBZ in_II effect_NN1 equivalent_NN1 to_II a_AT1 writer_NN1 's_GE underlining_VVG of_IO a_AT1 word_NN1 to_TO indicate_VVI emphasis_NN1 ._. 
There_EX is_VBZ a_AT1 sense_NN1 ,_, then_RT ,_, in_II which_DDQ the_AT analyst_NN1 is_VBZ creating_VVG the_AT text_NN1 which_DDQ others_NN2 will_VM read_VVI ._. 
In_II this_DD1 creation_NN1 of_IO the_AT written_JJ version_NN1 of_IO the_AT spoken_JJ text_NN1 he_PPHS1 makes_VVZ appeal_NN1 to_II conventional_JJ modes_NN2 of_IO interpretation_NN1 which_DDQ ,_, he_PPHS1 believes_VVZ ,_, are_VBR shared_VVN by_II other_JJ speakers_NN2 of_IO the_AT language_NN1 ._. 
It_PPH1 must_VM be_VBI further_RRR emphasised_VVN that_CST ,_, however_RGQV objective_JJ the_AT notion_NN1 of_IO '_GE text_NN1 '_GE may_VM appear_VVI as_CSA we_PPIS2 have_VH0 defined_VVN it_PPH1 (_( 'the_VV0 verbal_JJ record_NN1 of_IO a_AT1 communicative_JJ act'_NN1 )_) ,_, the_AT perception_NN1 and_CC interpretation_NN1 of_IO each_DD1 text_NN1 is_VBZ essentially_RR subjective_JJ ._. 
Different_JJ individuals_NN2 pay_VV0 attention_NN1 to_II different_JJ aspects_NN2 of_IO texts_NN2 ._. 
The_AT content_NN1 of_IO the_AT text_NN1 appeals_NN2 to_II them_PPHO2 or_CC fits_VVZ into_II their_APPGE experience_NN1 differently_RR ._. 
In_II discussing_VVG texts_NN2 we_PPIS2 idealise_VV0 away_II21 from_II22 this_DD1 variability_NN1 of_IO the_AT experiencing_NN1 of_IO the_AT text_NN1 and_CC assume_VV0 what_DDQ Schutz_NP1 has_VHZ called_VVN '_GE the_AT reciprocity_NN1 of_IO perspective_NN1 '_GE ,_, whereby_RRQ we_PPIS2 take_VV0 it_PPH1 for_IF granted_VVN that_CST readers_NN2 of_IO a_AT1 text_NN1 or_CC listeners_NN2 to_II a_AT1 text_NN1 share_VV0 the_AT same_DA experience_NN1 (_( Schutz_NP1 ,_, 1953_MC )_) ._. 
Clearly_RR for_IF a_AT1 great_JJ deal_NN1 of_IO ordinary_JJ everyday_JJ language_NN1 this_DD1 assumption_NN1 of_IO an_AT1 amount_NN1 of_IO overlap_NN1 of_IO point_NN1 of_IO view_NN1 sufficient_JJ to_TO allow_VVI mutual_JJ comprehension_NN1 is_VBZ necessary_JJ ._. 
From_RR41 time_RR42 to_RR43 time_RR44 however_RRQV we_PPIS2 are_VBR brought_VVN to_II a_AT1 halt_NN1 by_II different_JJ interpretations_NN2 of_IO '_GE the_AT same_DA text_NN1 '_GE ._. 
This_DD1 is_VBZ particularly_RR the_AT case_NN1 when_CS critical_JJ attention_NN1 is_VBZ being_VBG focused_VVN on_II details_NN2 of_IO spoken_JJ language_NN1 which_DDQ were_VBDR only_RR ever_RR intended_VVN by_II the_AT speaker_NN1 as_CSA ephemeral_JJ parts_NN2 ,_, relatively_RR unimportant_JJ ,_, of_IO the_AT working-out_NN1 of_IO what_DDQ he_PPHS1 wanted_VVD to_TO say_VVI ._. 
It_PPH1 seems_VVZ fair_JJ to_TO suggest_VVI that_DD1 discourse_NN1 analysis_NN1 of_IO spoken_JJ language_NN1 is_VBZ particularly_RR prone_JJ to_II overanalysis_NN1 ._. 
A_AT1 text_NN1 frequently_RR has_VHZ a_AT1 much_RR wider_JJR variety_NN1 of_IO interpretations_NN2 imposed_VVN upon_II it_PPH1 by_II analysts_NN2 studying_VVG it_PPH1 at_II their_APPGE leisure_NN1 ,_, than_CSN would_VM ever_RR have_VHI been_VBN possible_JJ for_IF the_AT participants_NN2 in_II the_AT communicative_JJ interaction_NN1 which_DDQ gives_VVZ rise_NN1 to_II the_AT '_GE text_NN1 '_GE ._. 
Once_CS the_AT analyst_NN1 has_58 '_GE created_JJ '_GE a_AT1 written_JJ transcription_NN1 from_II a_AT1 recorded_JJ spoken_JJ version_NN1 ,_, the_AT written_JJ text_NN1 is_VBZ available_JJ to_II him_PPHO1 in_RP just_RR the_AT way_NN1 a_AT1 literary_JJ text_NN1 is_VBZ available_JJ to_II the_AT literary_JJ critic_NN1 ._. 
It_PPH1 is_VBZ important_JJ to_TO remember_VVI ,_, when_CS we_PPIS2 discuss_VV0 spoken_JJ '_GE texts_NN2 '_GE ,_, the_AT transitoriness_NN1 of_IO the_AT original_JJ ._. 
It_PPH1 must_VM be_VBI clear_JJ that_CST our_APPGE simple_JJ definition_NN1 of_IO '_GE text_NN1 '_GE as_58 '_GE the_AT verbal_JJ record_NN1 of_IO a_AT1 communicative_JJ act_NN1 '_GE requires_VVZ at_RR21 least_RR22 two_MC hedges_NN2 :_: i_MC1 the_AT representation_NN1 of_IO a_AT1 text_NN1 which_DDQ is_VBZ presented_VVN for_IF discussion_NN1 may_VM in_RR21 part_RR22 ,_, particularly_RR where_CS the_AT written_JJ representation_NN1 of_IO a_AT1 spoken_JJ text_NN1 is_VBZ involved_VVN ,_, consist_VV0 of_IO a_AT1 prior_JJ analysis_NN1 (_( hence_RR interpretation_NN1 )_) of_IO a_AT1 fragment_NN1 of_IO discourse_NN1 by_II the_AT discourse_NN1 analyst_NN1 presenting_VVG the_AT text_NN1 for_IF consideration_NN1 ii_MC features_NN2 of_IO the_AT original_JJ production_NN1 of_IO the_AT language_NN1 ,_, for_REX21 example_REX22 shaky_JJ handwriting_NN1 or_CC quavering_VVG speech_NN1 ,_, are_VBR somewhat_RR arbitrarily_RR considered_VVN as_CSA features_NN2 of_IO the_AT text_NN1 rather_II21 than_II22 features_NN2 of_IO the_AT context_NN1 in_II which_DDQ the_AT language_NN1 is_VBZ produced_VVN ._. 
The_AT relationship_NN1 between_II speech_NN1 and_CC writing_VVG The_AT view_NN1 that_CST written_JJ language_NN1 and_CC spoken_JJ language_NN1 serve_VV0 ,_, in_RR21 general_RR22 ,_, quite_RG different_JJ functions_NN2 in_II society_NN1 has_VHZ been_VBN forcefully_RR propounded_VVN ,_, hardly_RR surprisingly_RR ,_, by_II scholars_NN2 whose_DDQGE main_JJ interest_NN1 lies_VVZ in_II anthropology_NN1 and_CC sociology_NN1 ._. 
Thus_RR Goody_JJ &amp;_CC Watt_NP1 (_( 1963_MC )_) and_CC Goody_NP1 (_( 1977_MC )_) suggest_VV0 that_CST analytic_JJ thinking_NN1 followed_VVD the_AT acquisition_NN1 of_IO written_JJ language_NN1 '_GE since_CS it_PPH1 was_VBDZ the_AT setting_NN1 down_RP of_IO speech_NN1 that_CST enabled_VVD man_NN1 clearly_RR to_TO separate_VVI words_NN2 ,_, to_TO manipulate_VVI their_APPGE order_NN1 and_CC to_TO develop_VVI syllogistic_JJ forms_NN2 of_IO reasoning_NN1 '_GE (_( Goody_NP1 ,_, 1977_MC :_: 11_MC )_) ._. 
Goody_NP1 goes_VVZ on_RP to_TO make_VVI even_RR larger_JJR claims_NN2 about_II the_AT ways_NN2 in_II which_DDQ the_AT acquisition_NN1 of_IO writing_NN1 ,_, which_DDQ permits_VVZ man_NN1 to_TO reflect_VVI upon_II what_DDQ he_PPHS1 has_VHZ thought_VVN ,_, has_VHZ permitted_VVN the_AT development_NN1 of_IO cognitive_JJ structures_NN2 which_DDQ are_VBR not_XX available_JJ to_II the_AT non-literate_JJ (_( cf._VV0 also_RR the_AT views_NN2 of_IO Vygotsky_NP1 ,_, 1962_MC )_) ._. 
He_PPHS1 examines_VVZ the_AT use_NN1 of_IO '_GE figures_NN2 of_IO the_AT written_JJ word_NN1 '_GE in_II various_JJ cultures_NN2 ,_, particularly_RR the_AT '_GE non-speech_JJ uses_NN2 of_IO language_NN1 '_GE which_DDQ develop_VV0 systems_NN2 of_IO classification_NN1 like_II lists_NN2 ,_, formulae_NN2 ,_, tables_NN2 and_CC '_GE recipes_NN2 for_IF the_AT organisation_NN1 and_CC development_NN1 of_IO human_JJ knowledge_NN1 '_GE (_( 1977_MC :_: 17_MC )_) ._. 
Goody_NP1 suggests_VVZ that_CST written_JJ language_NN1 has_VHZ two_MC main_JJ functions_NN2 :_: the_AT first_MD is_VBZ the_AT storage_NN1 function_NN1 which_DDQ permits_VVZ communication_NN1 over_II time_NNT1 and_CC space_NN1 ,_, and_CC the_AT second_NNT1 is_VBZ that_DD1 which_DDQ '_VBZ shifts_NN2 language_NN1 from_II the_AT oral_NN1 to_II the_AT visual_JJ domain'_NN1 and_CC permits_VVZ words_NN2 and_CC sentences_NN2 to_TO be_VBI examined_VVN out_II21 of_II22 their_APPGE original_JJ contexts_NN2 ,_, '_" where_RRQ they_PPHS2 appear_VV0 in_II a_AT1 very_RG different_JJ and_CC highly_RR '_GE abstract_NN1 '_GE context'_NN1 (_( 1977_MC :_: 78_MC )_) ._. 
It_PPH1 seems_VVZ reasonable_JJ to_TO suggest_VVI that_DD1 ,_, whereas_CS in_II daily_JJ life_NN1 in_II a_AT1 literate_JJ culture_NN1 ,_, we_PPIS2 use_VV0 speech_NN1 largely_RR for_IF the_AT establishment_NN1 and_CC maintenance_NN1 of_IO human_JJ relationships_NN2 (_( primarily_RR interactional_JJ use_NN1 )_) ,_, we_PPIS2 use_VV0 written_JJ language_NN1 largely_RR for_IF the_AT working_JJ out_II21 of_II22 and_CC transference_NN1 of_IO information_NN1 (_( primarily_RR transactional_JJ use_NN1 )_) ._. 
However_RR ,_, there_EX are_VBR occasions_NN2 when_RRQ speech_NN1 is_VBZ used_VVN for_IF the_AT detailed_JJ transmission_NN1 of_IO factual_JJ information_NN1 ._. 
It_PPH1 is_VBZ noteworthy_JJ ,_, then_RT ,_, that_CST the_AT recipient_NN1 often_RR writes_VVZ down_RP the_AT details_NN2 that_CST he_PPHS1 is_VBZ told_VVN ._. 
So_RR a_AT1 doctor_NN1 writes_VVZ down_RP his_APPGE patient_NN1 's_GE symptoms_NN2 ,_, an_AT1 architect_NN1 writes_VVZ down_RP his_APPGE client_NN1 's_GE requirements_NN2 ,_, Hansard_NP1 records_VVZ the_AT proceedings_NN2 of_IO the_AT British_JJ Parliament_NN1 ,_, we_PPIS2 write_VV0 down_RP friends_NN2 '_GE addresses_NN2 ,_, telephone_VV0 numbers_NN2 ,_, recipes_NN2 ,_, knitting_VVG patterns_NN2 ,_, and_RR31 so_RR32 on_RR33 ._. 
When_CS the_AT recipient_NN1 is_VBZ not_XX expected_VVN to_TO write_VVI down_RP the_AT details_NN2 ,_, it_PPH1 is_VBZ often_RR the_AT case_NN1 that_CST the_AT speaker_NN1 repeats_VVZ them_PPHO2 sometimes_RT several_DA2 times_NNT2 over_RP ._. 
Consider_VV0 the_AT typical_JJ structure_NN1 of_IO a_AT1 news_NN1 broadcast_NN1 which_DDQ opens_VVZ with_IW the_AT '_GE headlines_NN2 '_GE a_AT1 set_NN1 of_IO summary_NN1 statements_NN2 which_DDQ are_VBR followed_VVN by_II a_AT1 news_NN1 item_NN1 that_CST consists_VVZ of_IO an_AT1 expansion_NN1 and_CC repetition_NN1 of_IO the_AT first_MD headline_NN1 ,_, in_II which_DDQ is_VBZ embedded_VVN a_AT1 comment_NN1 from_II '_" our_APPGE man_NN1 on_II the_AT spot_NN1 '_GE that_DD1 recapitulates_VVZ the_AT main_JJ points_NN2 again_RT ,_, then_RT ,_, at_II the_AT end_NN1 of_IO the_AT broadcast_NN1 ,_, there_EX is_VBZ a_AT1 repetition_NN1 of_IO the_AT set_NN1 of_IO headlines_NN2 ._. 
There_EX is_VBZ a_AT1 general_JJ expectation_NN1 that_CST people_NN will_VM not_XX remember_VVI detailed_JJ facts_NN2 correctly_RR if_CS they_PPHS2 are_VBR only_RR exposed_VVN to_II them_PPHO2 in_II the_AT spoken_JJ mode_NN1 ,_, especially_RR if_CS they_PPHS2 are_VBR required_VVN to_TO remember_VVI them_PPHO2 over_RP an_AT1 extended_JJ period_NN1 of_IO time_NNT1 ._. 
This_DD1 aspect_NN1 of_IO communication_NN1 is_VBZ obviously_RR what_DDQ written_JJ language_NN1 is_VBZ supremely_RR good_JJ at_II ,_, whether_CSW for_IF the_AT benefit_NN1 of_IO the_AT individual_NN1 in_II remembering_VVG the_AT private_JJ paraphernalia_NN of_IO daily_JJ life_NN1 ,_, or_CC for_IF the_AT benefit_NN1 of_IO nations_NN2 in_II establishing_VVG constitutions_NN2 ,_, laws_NN2 and_CC treaties_NN2 with_IW other_JJ nations_NN2 ._. 
The_AT major_JJ differences_NN2 between_II speech_NN1 and_CC writing_NN1 derive_VV0 from_II the_AT fact_NN1 that_CST one_PN1 is_VBZ essentially_RR transitory_JJ and_CC the_AT other_NN1 is_VBZ designed_VVN to_TO be_VBI permanent_JJ ._. 
It_PPH1 is_VBZ exactly_RR this_DD1 point_NN1 which_DDQ D._NP1 J._NP1 Enright_NP1 makes_VVZ in_II the_AT observation_NN1 that_DD1 '_VBZ Plato_NP1 may_VM once_RR have_VHI thought_VVN more_RGR highly_RR of_IO speech_NN1 than_CSN of_IO writing_NN1 ,_, but_CCB I_PPIS1 doubt_VV0 he_PPHS1 does_VDZ now_RT !_! 
'_GE (_( Review_VV0 in_II The_AT Sunday_NPD1 Times_NNT2 ,_, 24_MC January_NPM1 1982_MC )_) ._. 
Differences_NN2 in_II form_NN1 between_II written_JJ and_CC spoken_JJ language_NN1 It_PPH1 is_VBZ not_XX our_APPGE intention_NN1 here_RL to_TO discuss_VVI the_AT many_DA2 different_JJ forms_NN2 of_IO spoken_JJ language_NN1 which_DDQ can_VM be_VBI identified_VVN even_RR within_II one_MC1 geographical_JJ area_NN1 like_II Britain_NP1 ._. 
Clearly_RR there_EX are_VBR dialectal_JJ differences_NN2 ,_, accent_NN1 differences_NN2 ,_, as_II31 well_II32 as_II33 '_GE register_NN1 '_GE differences_NN2 depending_II21 on_II22 variables_NN2 like_II the_AT topic_NN1 of_IO discussion_NN1 and_CC the_AT roles_NN2 of_IO the_AT participants_NN2 (_( see_VV0 e.g._REX Trudgill_NP1 ,_, 1974_MC and_CC Hudson_NP1 ,_, 1980_MC for_IF discussion_NN1 of_IO these_DD2 sorts_NN2 of_IO differences_NN2 )_) ._. 
There_EX is_VBZ however_RR ,_, one_MC1 further_JJR distinction_NN1 which_DDQ is_VBZ rarely_RR noted_VVN ,_, but_CCB which_DDQ it_PPH1 is_VBZ important_JJ to_TO draw_VVI attention_NN1 to_II here_RL ._. 
That_DD1 is_VBZ the_AT distinction_NN1 between_II the_AT speech_NN1 of_IO those_DD2 whose_DDQGE language_NN1 is_VBZ highly_RR influenced_VVN by_II long_JJ and_CC constant_JJ Immersion_NN1 in_II written_JJ language_NN1 forms_NN2 ,_, and_CC the_AT speech_NN1 of_IO those_DD2 whose_DDQGE language_NN1 is_VBZ relatively_RR uninfluenced_JJ by_II written_JJ forms_NN2 of_IO language_NN1 ._. 
It_PPH1 is_VBZ of_RR21 course_RR22 the_AT case_NN1 that_CST it_PPH1 is_VBZ the_AT speech_NN1 of_IO the_AT first_MD set_NN1 whose_DDQGE language_NN1 tends_VVZ to_TO be_VBI described_VVN in_II descriptions_NN2 of_IO the_AT language_NN1 (_( grammars_NN2 )_) ,_, since_CS descriptions_NN2 are_VBR typically_RR written_VVN by_II middle-aged_JJ people_NN who_PNQS have_VH0 spent_VVN long_JJ years_NNT2 reading_VVG written_JJ language_NN1 ._. 
In_RR21 particular_RR22 situations_NN2 the_AT speech_NN1 of_IO ,_, say_VV0 ,_, an_AT1 academic_JJ ,_, particularly_RR if_CS he_PPHS1 is_VBZ saying_VVG something_PN1 he_PPHS1 has_VHZ said_VVN or_CC thought_VVN about_II before_RT ,_, may_VM have_VHI a_AT1 great_JJ deal_NN1 in_II common_JJ with_IW written_JJ language_NN1 forms_NN2 ._. 
For_IF the_AT majority_NN1 of_IO the_AT population_NN1 ,_, even_RR of_IO a_AT1 '_GE literate_NN1 '_GE country_NN1 ,_, spoken_JJ language_NN1 will_VM have_VHI very_RG much_DA1 less_DAR in_II31 common_II32 with_II33 the_AT written_JJ language_NN1 ._. 
This_DD1 ,_, again_RT ,_, is_VBZ a_AT1 point_NN1 appreciated_VVN by_II Goody_NP1 :_: '_" Some_DD individuals_NN2 spend_VV0 more_DAR time_NNT1 with_IW the_AT written_JJ language_NN1 than_CSN they_PPHS2 do_VD0 with_IW the_AT spoken_JJ ._. 
Apart_II21 from_II22 the_AT effects_NN2 on_II their_APPGE own_DA personalities_NN2 ..._... what_DDQ are_VBR the_AT effects_NN2 on_II language_NN1 ?_? 
How_RRQ do_VD0 written_JJ languages_NN2 differ_VV0 from_II spoken_JJ ones_NN2 ?_? '_GE (_( 1977_MC :_: 124_MC )_) ._. 
In_II the_AT discussion_NN1 which_DDQ follows_VVZ we_PPIS2 shall_VM draw_VVI a_AT1 simplistic_JJ distinction_NN1 between_II spoken_JJ and_CC written_JJ language_NN1 which_DDQ takes_VVZ highly_RR literate_JJ written_JJ language_NN1 as_II the_AT norm_NN1 of_IO written_JJ language_NN1 ,_, and_CC the_AT speech_NN1 of_IO those_DD2 who_PNQS have_VH0 not_XX spent_VVN many_DA2 years_NNT2 exposed_VVN to_II written_JJ language_NN1 (_( a_AT1 set_NN1 which_DDQ will_VM include_VVI most_DAT young_JJ undergraduate_NN1 students_NN2 )_) as_II the_AT norm_NN1 for_IF spoken_JJ language_NN1 ._. 
In_II 1.2.1_MC we_PPIS2 discussed_VVD some_DD of_IO the_AT differences_NN2 in_II the_AT manner_NN1 of_IO production_NN1 of_IO speech_NN1 and_CC writing_NN1 ,_, differences_NN2 which_DDQ often_RR contribute_VV0 significantly_RR to_II characteristic_JJ forms_NN2 in_II written_JJ language_NN1 as_II21 against_II22 characteristic_JJ forms_NN2 in_II speech_NN1 ._. 
The_AT overall_JJ effect_NN1 is_VBZ to_TO produce_VVI speech_NN1 which_DDQ is_VBZ less_RGR richly_RR organised_VVN than_CSN written_JJ language_NN1 ,_, containing_VVG less_RGR densely_RR packed_JJ information_NN1 ,_, but_CCB containing_VVG more_RGR interactive_JJ markers_NN2 and_CC planning_NN1 '_GE fillers_NN2 '_GE ._. 
The_AT standard_JJ descriptive_JJ grammars_NN2 of_IO English_NN1 (_( e.g._REX Quirk_NN1 ,_, Greenbaum_NP1 ,_, Leech_NP1 &amp;_CC Svartvik_NP1 ,_, 1972_MC )_) typically_RR describe_VV0 features_NN2 of_IO the_AT written_JJ language_NN1 ,_, or_CC that_DD1 form_NN1 of_IO the_AT spoken_JJ language_NN1 which_DDQ is_VBZ highly_RR influenced_VVN by_II written_JJ language_NN1 ._. 
From_II the_AT descriptive_JJ work_NN1 of_IO a_AT1 number_NN1 of_IO scholars_NN2 studying_VVG spoken_JJ language_NN1 (_( e.g._REX Labov_NP1 ,_, 1972a_FO ;_; Sinclair_NP1 &amp;_CC Coulthard_NP1 ,_, 1975_MC ;_; Chafe_NP1 ,_, 1979_MC ;_; Ochs_NP1 ,_, 1979_MC ;_; Cicourel_NP1 ,_, 1981_MC ;_; Goffman_NP1 ,_, 1981_MC )_) we_PPIS2 can_VM extract_VVI some_DD (_( by_RR31 no_RR32 means_RR33 all_RR )_) features_NN2 which_DDQ characterise_VV0 spoken_JJ language_NN1 :_: a_AT1 the_AT syntax_NN1 of_IO spoken_JJ language_NN1 is_VBZ typically_RR much_RR less_RGR structured_JJ than_CSN that_DD1 of_IO written_JJ language_NN1 i_ZZ1 spoken_JJ language_NN1 contains_VVZ many_DA2 incomplete_JJ sentences_NN2 ,_, often_RR simply_RR sequences_NN2 ii_MC spoken_JJ language_NN1 typically_RR contains_VVZ rather_RG little_DA1 subordination_NN1 iii_MC in_II conversational_JJ speech_NN1 ,_, where_CS sentential_JJ syntax_NN1 can_VM be_VBI observed_VVN ,_, active_JJ declarative_JJ forms_NN2 are_VBR normally_RR found_VVN ._. 
In_RP over_RG 50_MC hours_NNT2 of_IO recorded_JJ conversational_JJ speech_NN1 ,_, Brown_NP1 ,_, Currie_NP1 and_CC Kenworthy_NP1 (_( 1980_MC )_) found_VVD very_RG few_DA2 examples_NN2 of_IO passives_NN2 ,_, it-clefts_NN2 or_CC wh-clefts_NN2 ._. 
Crystal(1980)_FO also_RR presents_VVZ some_DD of_IO the_AT problems_NN2 encountered_VVN in_II attempting_VVG to_TO analyse_VVI spontaneous_JJ speech_NN1 in_II31 terms_II32 of_II33 categories_NN2 like_II sentence_NN1 and_CC clause_NN1 As_II a_AT1 brief_JJ example_NN1 ,_, notice_VV0 how_RRQ this_DD1 speaker_NN1 pauses_NN2 and_CC begins_VVZ each_DD1 new_JJ '_GE sentence_NN1 '_GE before_II formally_RR completing_VVG the_AT previous_JJ one_PN1 :_: it_PPH1 's_VBZ quite_RG nice_RR the_AT Grassmarket_NN1 since_CS +_FO it_PPH1 's_VHZ always_RR had_VHN the_AT antique_JJ shops_NN2 but_CCB they_PPHS2 're_VBR looking_VVG +_FO they_PPHS2 're_VBR sort_RR21 of_RR22 +_FO em_FU +_FO become_VV0 a_RR21 bit_RR22 nicer_JJR +_FO b_ZZ1 in_II written_JJ language_NN1 an_AT1 extensive_JJ set_NN1 of_IO metalingual_JJ markers_NN2 exists_VVZ to_TO mark_VVI relationships_NN2 between_II clauses_NN2 (_( that_CST complementisers_NN2 ,_, when_CS I_PPIS1 while_VV0 temporal_JJ markers_NN2 ,_, so-called_JJ '_GE logical_JJ connectors_NN2 '_GE like_JJ besides_RR ,_, moreover_RR ,_, however_RR ,_, in_II31 spite_II32 of_II33 ,_, etc._RA )_) ,_, in_II spoken_JJ language_NN1 the_AT largely_RR paratactically_RR organised_JJ chunks_NN2 are_VBR related_VVN by_II and_CC ,_, but_CCB ,_, then_RT and_CC ,_, more_RGR rarely_RR ,_, if_CS ._. 
The_AT speaker_NN1 is_VBZ typically_RR less_RGR explicit_JJ than_CSN the_AT writer_NN1 :_: I_PPIS1 'm_VBM so_RG tired_JJ (_( because_CS )_) I_PPIS1 had_VHD to_TO walk_VVI all_DB the_AT way_NN1 home_RL ._. 
In_II written_JJ language_NN1 rhetorical_JJ organisers_NN2 of_IO larger_JJR stretches_NN2 of_IO discourse_NN1 appear_VV0 ,_, like_II firstly_RR ,_, more_RGR important_JJ than_CSN and_CC in_II conclusion_NN1 ._. 
These_DD2 are_VBR rare_JJ in_II spoken_JJ language._NNU c_ZZ1 In_II written_JJ language_NN1 ,_, rather_RG heavily_RR premodified_JJ noun_NN1 phrases_NN2 (_( like_II that_DD1 one_MC1 )_) are_VBR quite_RG common_JJ it_PPH1 is_VBZ rare_JJ in_II spoken_JJ language_NN1 to_TO find_VVI more_DAR than_CSN two_MC premodifying_JJ adjectives_NN2 and_CC there_EX is_VBZ a_AT1 strong_JJ tendency_NN1 to_II structure_NN1 the_AT short_JJ chunks_NN2 of_IO speech_NN1 so_CS21 that_CS22 only_RR one_MC1 predicate_NN1 is_VBZ attached_VVN to_II a_AT1 given_JJ referent_NN1 at_II a_AT1 time_NNT1 (_( simple_JJ case-frame_NN1 or_CC one-place_JJ predicate_NN1 )_) as_CSA in_II :_: it_PPH1 's_VBZ a_AT1 biggish_JJ cat_NN1 +_FO tabby_NN1 +_FO with_IW torn_JJ ears_NN2 ,_, or_CC in_II :_: old_JJ man_NN1 McArthur_NP1 +_FO he_PPHS1 was_VBDZ a_AT1 wee_JJ chap_NN1 +_FO oh_UH very_RG small_JJ +_FO and_CC eh_UH a_AT1 beard_NN1 +_FO and_CC he_PPHS1 was_VBDZ pretty_RG stooped_JJ ._. 
The_AT packaging_NN1 of_IO information_NN1 related_VVN to_II a_AT1 particular_JJ referent_NN1 can_VM ,_, in_II the_AT written_JJ language_NN1 ,_, be_VBI very_RG concentrated_JJ ,_, as_CSA in_II the_AT following_JJ news_NN1 item_NN1 :_: A_AT1 man_NN1 who_PNQS turned_VVD into_II a_AT1 human_JJ torch_NN1 ten_MC days_NNT2 ago_RA after_II snoozing_VVG in_II his_APPGE locked_JJ car_NN1 while_CS smoking_VVG his_APPGE pipe_NN1 has_VHZ died_VVN in_II hospital_NN1 ._. 
(_( Evening_NNT1 News_NN1 (_( Edinburgh_NP1 )_) ,_, 22_MC April_NPM1 1982_MC )_) d_ZZ1 Whereas_CS written_JJ language_NN1 sentences_NN2 are_VBR generally_RR structured_VVN in_II subject-predicate_JJ form_NN1 ,_, in_II spoken_JJ language_NN1 it_PPH1 is_VBZ quite_RG common_JJ to_TO find_VVI what_DDQ Givo/1n_FU (_( 1979b_FO )_) calls_VVZ topic-comment_JJ structure_NN1 ,_, as_CSA in_II the_AT cats_NN2 +_FO did_VDD you_PPY let_VVI them_PPHO2 out._NNU e_ZZ1 in_II informal_JJ speech_NN1 ,_, the_AT occurrence_NN1 of_IO passive_JJ constructions_NN2 is_VBZ relatively_RR infrequent_JJ ._. 
That_DD1 use_NN1 of_IO the_AT passive_JJ in_II written_JJ language_NN1 which_DDQ allows_VVZ non-attribution_NN1 of_IO agency_NN1 is_VBZ typically_RR absent_JJ from_II conversational_JJ speech_NN1 ._. 
Instead_RR ,_, active_JJ constructions_NN2 with_IW indeterminate_JJ group_NN1 agents_NN2 are_VBR noticeable_JJ ,_, as_CSA in_II :_: Oh_UH everything_PN1 they_PPHS2 do_VD0 in_II Edinburgh_NP1 +_FO they_PPHS2 do_VD0 it_PPH1 far_RR too_RG slowly_RR f_ZZ1 in_II chat_NN1 about_II the_AT immediate_JJ environment_NN1 ,_, the_AT speaker_NN1 may_VM rely_VVI on_II (_( e.g._REX )_) gaze_NN1 direction_NN1 to_TO supply_VVI a_AT1 referent_NN1 :_: (_( looking_VVG at_II the_AT rain_NN1 )_) frightful_JJ is_VBZ n't_XX it._NNU g_ZZ1 the_AT speaker_NN1 may_VM replace_VVI or_CC refine_VVI expressions_NN2 as_CSA he_PPHS1 goes_VVZ along_RP :_: this_DD1 man_NN1 +_FO this_DD1 chap_NN1 she_PPHS1 was_VBDZ going_VVG out_RP with_IW h_ZZ1 the_AT speaker_NN1 typically_RR uses_VVZ a_AT1 good_JJ deal_NN1 of_IO rather_RG generalised_JJ vocabulary_NN1 :_: a_AT1 lot_NN1 of_IO ,_, got_VVD ,_, do_VD0 ,_, thing_NN1 ,_, nice_JJ ,_, stuff_NN1 ,_, place_NN1 and_CC things_NN2 like_II that._DD1 i_ZZ1 the_AT speaker_NN1 frequently_RR repeats_VVZ the_AT same_DA syntactic_JJ form_NN1 several_DA2 times_NNT2 over_RP ,_, as_CSA this_DD1 fairground_NN1 inspector_NN1 does_VDZ :_: I_PPIS1 look_VV0 at_II fire_NN1 extinguishers_NN2 +_FO I_MC1 look_NN1 at_II fire_NN1 exits_VVZ +_FO I_MC1 look_NN1 at_II what_DDQ gangways_NN2 are_VBR available_JJ +_FO I_MC1 look_NN1 at_II electric_JJ cables_NN2 what_DDQ +_FO are_VBR they_PPHS2 properly_RR earthed_VVN +_FO are_VBR they_PPHS2 properly_RR covered_VVN j_ZZ1 the_AT speaker_NN1 may_VM produce_VVI a_AT1 large_JJ number_NN1 of_IO prefabricated_JJ '_GE fillers_NN2 '_GE :_: well_RR ,_, erm_FU ,_, I_PPIS1 think_VV0 ,_, you_PPY know_VV0 ,_, if_CS you_PPY see_VV0 what_DDQ I_PPIS1 mean_VV0 ,_, of_RR21 course_RR22 ,_, and_RR31 so_RR32 on_RR33 ._. 
Some_DD of_IO the_AT typical_JJ distinctions_NN2 between_II discourse_NN1 which_DDQ has_VHZ been_VBN written_VVN and_CC that_DD1 which_DDQ has_VHZ been_VBN spoken_VVN can_VM be_VBI seen_VVN in_II the_AT following_JJ two_MC descriptions_NN2 of_IO a_AT1 rainbow_NN1 ._. 
(_( No_AT direct_JJ comparison_NN1 is_VBZ intended_VVN ,_, since_CS the_AT two_MC pieces_NN2 of_IO discourse_NN1 were_VBDR produced_VVN in_II strictly_RR non-comparable_JJ circumstances_NN2 for_IF very_RG different_JJ purposes_NN2 ._. )_) 
And_CC then_RT ,_, in_II the_AT blowing_NN1 clouds_NN2 ,_, she_PPHS1 saw_VVD a_AT1 band_NN1 of_IO faint_JJ iridescence_NN1 colouring_VVG in_II faint_JJ shadows_NN2 a_AT1 portion_NN1 of_IO the_AT hill_NN1 ._. 
And_CC forgetting_VVG ,_, startled_VVD ,_, she_PPHS1 looked_VVD for_IF the_AT hovering_JJ colour_NN1 and_CC saw_VVD a_AT1 rainbow_NN1 forming_VVG itself_PPX1 ._. 
In_II one_MC1 place_VV0 it_PPH1 gleamed_VVD fiercely_RR ,_, and_CC ,_, her_APPGE heart_NN1 anguished_JJ with_IW hope_NN1 ,_, she_PPHS1 sought_VVD the_AT shadow_NN1 of_IO iris_NN1 where_CS the_AT bow_NN1 should_VM be_VBI ._. 
Steadily_RR the_AT colour_NN1 gathered_VVD ,_, mysteriously_RR ,_, from_II nowhere_RL ,_, it_PPH1 took_VVD presence_NN1 upon_II itself_PPX1 ,_, there_EX was_VBDZ a_AT1 faint_JJ ,_, vast_JJ rainbow_NN1 ._. 
(_( D._NP1 H._NP1 Lawrence_NP1 ,_, The_AT Rainbow_NN1 ,_, chapter_NN1 16_MC )_) In_II the_AT first_MD extract_NN1 (_( I_ZZ1 )_) ,_, the_AT rich_JJ lexis_NN1 and_CC well-organised_JJ structure_NN1 are_VBR indications_NN2 that_CST the_AT writer_NN1 has_VHZ taken_VVN time_NNT1 in_II the_AT construction_NN1 ,_, and_CC possibly_RR reconstruction_NN1 after_II several_DA2 rewritings_NN2 ,_, of_IO the_AT final_JJ product_NN1 ._. 
There_EX are_VBR complete_JJ sentences_NN2 ,_, containing_VVG subordinations_NN2 ,_, frequent_JJ modifications_NN2 via_II adjectives_NN2 and_CC adverbs_NN2 ,_, and_CC more_DAR than_CSN single_JJ predicates_NN2 per_II referential_JJ expression_NN1 ._. 
In_II extract_NN1 (_( 2_MC )_) ,_, there_EX are_VBR frequent_JJ pauses_NN2 ,_, often_RR interrupting_VVG major_JJ syntactic_JJ units_NN2 ,_, repetitions_NN2 ,_, incomplete_JJ sentences_NN2 ,_, generalised_JJ vocabulary_NN1 ,_, fillers_NN2 and_CC one_MC1 example_NN1 of_IO a_AT1 tongue-slip._JJ normally_RR after_CS +_FO very_RG heavy_JJ rain_NN1 +_FO or_CC something_PN1 like_II that_DD1 +_FO and_CC +_FO you_PPY 're_VBR driving_VVG along_II the_AT road_NN1 +_FO and_CC +_FO far_RR away_RL +_FO you_PPY see_VV0 +_FO well_RR +_FO er_FU +_FO a_AT1 series_NN +_FO of_IO +_FO stripes_NN2 +_FO +_FO formed_VVN like_II a_AT1 bow_NN1 +_FO an_AT1 arch_NN1 +_FO +_FO very_RG very_RG far_RR away_RL +_FO ah_UH +_FO seven_MC colours_NN2 but_CCB +_FO +_FO I_ZZ1 guess_VV0 you_PPY hardly_RR ever_RR see_VV0 seven_MC it_PPH1 's_VBZ just_RR a_ZZ1 +_FO a_AT1 series_NN of_IO +_FO colours_NN2 which_DDQ +_FO they_PPHS2 seem_VV0 to_TO be_VBI separate_JJ but_CCB if_CS you_PPY try_VV0 to_TO look_VVI for_IF the_AT separate_JJ (_( kAz_NN1 )_) colours_NN2 they_PPHS2 always_RR seem_VV0 +_FO very_RG hard_RR +_FO to_TO separate_VVI +_FO if_CS you_PPY see_VV0 what_DDQ I_PPIS1 mean_VV0 ++_FO (_( Postgraduate_NN1 student_NN1 speaking_VVG informally_RR )_) The_AT speaker_NN1 planning_VVG in_II the_AT here-and-now_RT ,_, possibly_RR threatened_VVN with_IW his_APPGE interlocutor_NN1 wanting_VVG to_TO take_VVI a_AT1 turn_NN1 ,_, typically_RR repeats_VVZ himself_PPX1 a_RR31 good_RR32 deal_RR33 ,_, using_VVG the_AT same_DA syntactic_JJ structure_NN1 ,_, the_AT same_DA lexical_JJ items_NN2 ,_, using_VVG the_AT first_MD word_NN1 that_CST comes_VVZ to_TO mind_VVI rather_II21 than_II22 hunting_NN1 for_IF the_AT mot_NN121 juste_NN122 ,_, filling_VVG in_II pauses_NN2 with_IW '_GE fillers_NN2 '_GE ._. 
The_AT overall_JJ effect_NN1 is_VBZ of_IO information_NN1 produced_VVN in_II a_AT1 much_RR less_RGR dense_JJ manner_NN1 than_CSN is_VBZ characteristic_JJ of_IO written_JJ language_NN1 ._. 
We_PPIS2 must_VM assume_VVI that_CST the_AT density_NN1 of_IO information_NN1 packing_VVG in_II spoken_JJ language_NN1 is_VBZ appropriate_JJ for_IF the_AT listener_NN1 to_TO process_VVI comfortably_RR ._. 
Most_DAT people_NN have_VH0 experienced_VVN expository_JJ prose_NN1 read_VVD aloud_RR which_DDQ they_PPHS2 have_VH0 found_VVN difficult_JJ to_TO follow_VVI in_II the_AT spoken_JJ mode_NN1 ._. 
Few_DA2 people_NN can_VM extract_VVI a_RR31 great_RR32 deal_RR33 from_II a_AT1 lecture_NN1 which_DDQ is_VBZ read_VVN aloud_RR with_IW no_AT visual_JJ support_NN1 ._. 
Goody_NP1 points_VVZ out_RP that_CST the_AT written_JJ form_NN1 of_IO language_NN1 releases_VVZ us_PPIO2 from_II the_AT linear_JJ experiential_JJ mode_NN1 :_: '_GE the_AT fact_NN1 that_CST it_PPH1 takes_VVZ a_AT1 visual_JJ form_NN1 means_VVZ that_CST one_PN1 can_VM escape_VVI from_II the_AT problem_NN1 of_IO the_AT succession_NN1 of_IO events_NN2 in_II time_NNT1 ,_, by_II backtracking_VVG ,_, skipping_VVG ,_, looking_VVG to_TO see_VVI who-done-it_PPH1 before_CS we_PPIS2 know_VV0 what_DDQ it_PPH1 is_VBZ they_PPHS2 did_VDD ._. 
Who_PNQS ,_, except_CS the_AT most_RGT obsessive_JJ academic_JJ ,_, reads_VVZ a_AT1 book_NN1 as_CSA he_PPHS1 hears_VVZ speech_NN1 ?_? 
Who_PNQS ,_, except_CS the_AT most_RGT avant-garde_JJ of_IO modern_JJ dramatists_NN2 ,_, attempts_NN2 to_TO write_VVI as_CSA they_PPHS2 speak_VV0 ?_? '_GE (_( 1977:124_MC )_) ._. 
Sentence_NN1 and_CC utterance_NN1 It_PPH1 might_VM seem_VVI reasonable_JJ to_TO propose_VVI that_CST the_AT features_NN2 of_IO spoken_JJ language_NN1 outlined_VVN in_II the_AT preceding_JJ section_NN1 should_VM be_VBI considered_VVN as_CSA features_NN2 of_IO utterances_NN2 ,_, and_CC those_DD2 features_NN2 typical_JJ of_IO written_JJ language_NN1 as_CSA characteristic_NN1 of_IO sentences_NN2 ._. 
In_II this_DD1 convenient_JJ distinction_NN1 ,_, we_PPIS2 can_VM say_VVI ,_, in_II a_AT1 fairly_RR non-technical_JJ way_NN1 ,_, that_CST utterances_NN2 are_VBR spoken_VVN and_CC sentences_NN2 are_VBR written_VVN and_CC that_CST we_PPIS2 will_VM apply_VVI these_DD2 terms_NN2 to_II what_DDQ Lyons_NP1 describes_VVZ as_II '_GE the_AT products_NN2 of_IO ordinary_JJ language-behaviour_NN1 '_GE ._. 
In_II the_AT case_NN1 of_IO the_AT term_NN1 sentence_NN1 ,_, it_PPH1 is_VBZ important_JJ to_TO be_VBI clear_JJ about_II the_AT type_NN1 of_IO object_NN1 one_PN1 is_VBZ referring_VVG to_II ._. 
Lyons_NP1 makes_VVZ a_AT1 distinction_NN1 between_RL '_GE text-sentences_NN2 '_GE and_CC '_GE system-sentences_NN2 '_GE ._. 
He_PPHS1 describes_VVZ the_AT latter_DA in_II the_AT following_JJ way_NN1 :_: system-sentences_NN2 never_RR occur_VV0 as_II the_AT products_NN2 of_IO ordinary_JJ language-behaviour_NN1 ._. 
Representations_NN2 of_IO system-sentences_NN2 may_VM of_RR21 course_RR22 be_VBI used_VVN in_II metalinguistic_JJ discussion_NN1 of_IO the_AT structure_NN1 and_CC functions_NN2 of_IO language_NN1 :_: and_CC it_PPH1 is_VBZ such_DA representations_NN2 that_CST are_VBR customarily_RR cited_VVN in_II grammatical_JJ descriptions_NN2 of_IO particular_JJ languages_NN2 ._. 
(_( Lyons_NP1 ,_, 1977:31_MC )_) Since_CS the_AT linguistic_JJ exemplification_NN1 presented_VVN in_II31 support_II32 of_II33 our_APPGE discussion_NN1 throughout_II this_DD1 book_NN1 is_VBZ overwhelmingly_RR drawn_VVN from_II '_GE ordinary_JJ language_NN1 behaviour_NN1 '_GE ,_, we_PPIS2 shall_VM generally_RR employ_VVI the_AT term_NN1 '_GE sentence_NN1 '_GE in_II the_AT '_GE text-sentence_NN1 '_GE ,_, and_CC not_XX the_AT '_GE system-sentence_NN1 '_GE sense_NN1 ._. 
Although_CS the_AT linguist_NN1 who_PNQS undertakes_VVZ the_AT analysis_NN1 of_IO discourse_NN1 has_VHZ ultimately_RR the_AT same_DA aims_NN2 as_CSA a_AT1 linguist_NN1 who_PNQS uses_NN2 '_GE system-sentences_NN2 '_GE in_II his_APPGE grammatical_JJ description_NN1 of_IO a_AT1 language_NN1 ,_, there_EX are_VBR important_JJ methodological_JJ differences_NN2 involved_JJ in_II the_AT two_MC approaches_NN2 ._. 
Both_DB2 linguists_NN2 wish_VV0 to_TO produce_VVI accurate_JJ descriptions_NN2 of_IO the_AT particular_JJ language_NN1 studied_VVD ._. 
In_II31 pursuit_II32 of_II33 this_DD1 goal_NN1 ,_, the_AT grammarian_NN1 will_VM concentrate_VVI on_II a_AT1 particular_JJ body_NN1 of_IO data_NN and_CC attempt_VV0 to_TO produce_VVI an_AT1 exhaustive_JJ but_CCB economical_JJ set_NN1 of_IO rules_NN2 which_DDQ will_VM account_VVI for_IF all_DB and_CC only_RR the_AT acceptable_JJ sentences_NN2 in_II his_APPGE data_NN ._. 
He_PPHS1 will_VM not_XX normally_RR seek_VVI to_TO account_VVI for_IF the_AT mental_JJ processes_NN2 involved_JJ in_II any_DD language-user_NN1 's_GE production_NN1 of_IO those_DD2 sentences_NN2 ,_, nor_CC to_TO describe_VVI the_AT physical_JJ or_CC social_JJ contexts_NN2 in_II which_DDQ those_DD2 sentences_NN2 occur_VV0 ._. 
On_II each_DD1 of_IO these_DD2 issues_NN2 ,_, concerning_VVG '_GE data_NN '_GE ,_, '_GE rules_NN2 '_GE ,_, '_GE processes_NN2 '_GE and_CC '_GE contexts_NN2 '_GE ,_, the_AT discourse_NN1 analyst_NN1 will_VM take_VVI a_AT1 different_JJ view_NN1 ._. 
On_II '_GE data'_NN1 The_AT grammarian_NN1 's_VBZ '_GE data_NN '_GE is_VBZ inevitably_RR the_AT single_JJ sentence_NN1 ,_, or_CC a_AT1 set_NN1 of_IO single_JJ sentences_NN2 illustrating_VVG a_AT1 particular_JJ feature_NN1 of_IO the_AT language_NN1 being_VBG studied_VVN ._. 
It_PPH1 is_VBZ also_RR typically_RR the_AT case_NN1 that_CST the_AT grammarian_NN1 will_VM have_VHI constructed_VVN the_AT sentence_NN1 or_CC sentences_NN2 he_PPHS1 uses_VVZ as_CSA examples_NN2 ._. 
This_DD1 procedure_NN1 is_VBZ not_XX often_RR made_VVN explicit_JJ ,_, but_CCB an_AT1 overt_JJ commitment_NN1 to_II the_AT constructed-data_JJ approach_NN1 has_VHZ recently_RR been_VBN expressed_VVN in_II the_AT following_JJ terms_NN2 :_: I_PPIS1 shall_VM assume_VVI ..._... that_CST invented_VVD strings_NN2 and_CC certain_JJ intuitive_JJ judgements_NN2 about_II them_PPHO2 constitute_VV0 legitimate_JJ data_NN for_IF linguistic_JJ research_NN1 ._. 
(_( Gazdar_NN1 ,_, 1979_MC :_: 11_MC )_) In_II contrast_NN1 ,_, the_AT analysis_NN1 of_IO discourse_NN1 ,_, as_CSA undertaken_VVN and_CC exemplified_VVN in_II this_DD1 book_NN1 ,_, is_VBZ typically_RR based_VVN on_II the_AT linguistic_JJ output_NN1 of_IO someone_PN1 other_II21 than_II22 the_AT analyst_NN1 ._. 
On_II the_AT few_DA2 occasions_NN2 where_RRQ constructed_JJ data_NN is_VBZ used_VVN as_CSA illustration_NN1 (_( of_IO a_AT1 paradigm_NN1 ,_, for_REX21 example_REX22 ,_, in_II Chapter_NN1 4_MC )_) ,_, it_PPH1 is_VBZ inevitably_RR directed_VVN towards_II accounting_VVG for_IF the_AT range_NN1 of_IO formal_JJ options_NN2 available_JJ to_II a_AT1 speaker_NN1 or_CC writer_NN1 ._. 
More_RGR typically_RR ,_, the_AT discourse_NN1 analyst_NN1 's_VBZ '_GE data_NN '_GE is_VBZ taken_VVN from_II written_JJ texts_NN2 or_CC tape-recordings_NN2 ._. 
It_PPH1 is_VBZ rarely_RR in_II the_AT form_NN1 of_IO a_AT1 single_JJ sentence_NN1 ._. 
This_DD1 type_NN1 of_IO linguistic_JJ material_NN1 is_VBZ sometimes_RT described_VVN as_II '_GE performance-data_NN '_GE and_CC may_VM contain_VVI features_NN2 such_II21 as_II22 hesitations_NN2 ,_, slips_NN2 ,_, and_CC non-standard_JJ forms_NN2 which_DDQ a_AT1 linguist_NN1 like_II Chomsky_NP1 (_( 1965_MC )_) believed_VVN should_VM not_XX have_VHI to_TO be_VBI accounted_VVN for_IF in_II the_AT grammar_NN1 of_IO a_AT1 language_NN1 ._. 
Although_CS these_DD2 two_MC views_NN2 of_IO '_GE data_NN '_GE differ_VV0 substantially_RR ,_, they_PPHS2 are_VBR not_XX incompatible_JJ ,_, unless_CS they_PPHS2 are_VBR taken_VVN in_II an_AT1 extreme_JJ form_NN1 ._. 
A_AT1 discourse_NN1 analyst_NN1 may_VM regularly_RR work_VVI with_IW extended_JJ extracts_NN2 of_IO conversational_JJ speech_NN1 ,_, for_REX21 example_REX22 ,_, but_CCB he_PPHS1 does_VDZ not_XX consider_VVI his_APPGE data_NN in_II isolation_NN1 from_II the_AT descriptions_NN2 and_CC insights_NN2 provided_VVN by_II sentence-grammarians_NN2 ._. 
It_PPH1 should_VM be_VBI the_AT case_NN1 that_CST a_AT1 linguist_NN1 who_PNQS is_VBZ primarily_RR interested_JJ in_II the_AT analysis_NN1 of_IO discourse_NN1 is_VBZ ,_, in_II some_DD sense_NN1 ,_, also_RR a_AT1 sentence-grammarian_NN1 ._. 
Similarly_RR ,_, the_AT sentence-grammarian_NN1 can_VM not_XX remain_VVI immured_VVN from_II the_AT discourse_NN1 he_PPHS1 encounters_VVZ in_II his_APPGE daily_JJ life_NN1 ._. 
The_AT sentence_NN1 he_PPHS1 constructs_NN2 to_TO illustrate_VVI a_AT1 particular_JJ linguistic_JJ feature_NN1 must_VM ,_, in_II some_DD sense_NN1 ,_, derive_VV0 from_II the_AT '_GE ordinary_JJ language_NN1 '_GE of_IO his_APPGE daily_JJ life_NN1 and_CC also_RR be_VBI acceptable_JJ in_II it_PPH1 ._. 
A_AT1 dangerously_RR extreme_JJ view_NN1 of_IO '_GE relevant_JJ data_NN '_GE for_IF a_AT1 discourse_NN1 analyst_NN1 would_VM involve_VVI denying_VVG the_AT admissibility_NN1 of_IO a_AT1 constructed_JJ sentence_NN1 as_CSA linguistic_JJ data_NN ._. 
Another_DD1 would_VM be_VBI an_AT1 analytic_JJ approach_NN1 to_II data_NN which_DDQ did_VDD not_XX require_VVI that_CST there_EX should_VM be_VBI linguistic_JJ evidence_NN1 in_II the_AT data_NN to_TO support_VVI analytic_JJ claims_NN2 ._. 
We_PPIS2 shall_VM return_VVI to_II the_AT issue_NN1 of_IO '_GE relevant_JJ data_NN '_GE for_IF discourse_NN1 analysis_NN1 in_II Chapter_NN1 2_MC ._. 
An_AT1 over-extreme_JJ view_NN1 of_IO what_DDQ counts_VVZ as_CSA data_NN for_IF the_AT sentence-grammarian_NN1 was_VBDZ ,_, according_II21 to_II22 Sampson_NP1 (_( 1980_MC )_) ,_, noticeable_JJ in_II some_DD of_IO the_AT early_JJ work_NN1 of_IO generative_JJ grammarians_NN2 ._. 
Chomsky_NP1 gave_VVD an_AT1 indication_NN1 of_IO the_AT narrowness_NN1 of_IO view_NN1 which_DDQ could_VM be_VBI taken_VVN ,_, when_RRQ ,_, immediately_RR before_II his_APPGE conclusion_NN1 that_DD1 '_VBZ grammar_NN1 is_VBZ autonomous_58 '_GE ,_, he_PPHS1 stated_VVD :_: Despite_II the_AT undeniable_JJ interest_NN1 and_CC importance_NN1 of_IO semantic_JJ and_CC statistical_JJ studies_NN2 of_IO language_NN1 ,_, they_PPHS2 appear_VV0 to_TO have_VHI no_AT direct_JJ relevance_NN1 to_II the_AT problem_NN1 of_IO determining_JJ or_CC characterising_VVG the_AT set_NN1 of_IO grammatical_JJ utterances_NN2 ._. 
(_( Chomsky_NP1 ,_, 1957_MC :_: 17_MC )_) The_AT essential_JJ problem_NN1 in_II an_AT1 extreme_JJ version_NN1 of_IO the_AT constructed-sentence_JJ approach_NN1 occurs_VVZ when_RRQ the_AT resulting_JJ sentences_NN2 are_VBR tested_VVN only_RR against_II the_AT linguist_NN1 's_GE introspection_NN1 ._. 
This_DD1 can_VM (_( and_CC occasionally_RR did_VDD )_) lead_VV0 to_II a_AT1 situation_NN1 in_II which_DDQ a_AT1 linguist_NN1 claims_VVZ that_CST the_AT '_GE data_NN '_GE he_PPHS1 is_VBZ using_VVG illustrates_VVZ acceptable_JJ linguistic_JJ strings_NN2 because_CS he_PPHS1 says_VVZ it_PPH1 does_VDZ ,_, as_II a_AT1 result_NN1 of_IO personal_JJ introspection_NN1 ,_, and_CC regardless_RR of_IO how_RGQ many_DA2 voices_NN2 arise_VV0 in_II disagreement_NN1 ._. 
The_AT source_NN1 of_IO this_DD1 problem_NN1 ,_, as_CSA Sampson_NP1 (_( 1980_MC :_: 153_MC )_) points_VVZ out_RP ,_, is_VBZ that_DD1 the_AT narrow_JJ restriction_NN1 of_IO '_GE data_NN '_GE to_II constructed_JJ sentences_NN2 and_CC personal_JJ introspection_NN1 leads_VVZ to_II a_AT1 '_GE non-testability_NN1 '_GE ,_, in_II principle_NN1 ,_, of_IO any_DD claims_NN2 made_VVD ._. 
One_MC1 outcome_NN1 of_IO this_DD1 narrow_JJ view_NN1 of_IO data_NN is_VBZ that_CST there_EX is_VBZ a_AT1 concentration_NN1 on_II '_GE artificially_RR contrived_JJ sentences_NN2 isolated_VVN from_II their_APPGE communicative_JJ context_NN1 '_GE (_( see_VV0 Preface_NN1 to_II Givn_NP1 (_( ed._NN1 )_) ,_, 1979_MC )_) ._. 
Although_CS we_PPIS2 shall_VM appeal_VVI frequently_RR ,_, in_II the_AT course_NN1 of_IO this_DD1 book_NN1 ,_, to_II the_AT insights_NN2 of_IO sentence_NN1 grammarians_NN2 ,_, including_II those_DD2 working_VVG within_II a_AT1 generative_JJ framework_NN1 ,_, we_PPIS2 shall_VM avoid_VVI as_RG far_RR as_CSA possible_JJ the_AT methodology_NN1 which_DDQ depends_VVZ on_II what_DDQ Lyons_NP1 (_( I968_FO )_) described_VVD as_CSA regularised_VVN ,_, standardised_JJ and_CC decontextualised_JJ data_NN ._. 
Rules_NN2 versus_II regularities_NN2 A_ZZ1 corollary_NN1 to_II the_AT restricted_JJ data_NN approach_NN1 found_VVN in_II much_DA1 of_IO Chomskyan_JJ linguistics_NN1 is_VBZ the_AT importance_NN1 placed_VVN on_II writing_VVG rules_NN2 of_IO grammar_NN1 which_DDQ are_VBR fixed_VVN and_CC true_JJ 100%_NNU of_IO the_AT time_NNT1 ._. 
Just_RR as_CSA the_AT grammarian_NN1 's_VBZ '_GE data_NN '_GE can_NN1 not_XX contain_VVI any_DD variable_JJ phenomena_NN2 ,_, so_CS the_AT grammar_NN1 must_VM have_VHI categorial_JJ rules_NN2 ,_, and_CC not_XX '_GE rules_NN2 '_GE which_DDQ are_VBR true_JJ only_RR some_DD of_IO the_AT time_NNT1 ._. 
It_PPH1 is_VBZ typical_JJ of_IO arguments_NN2 concerning_II the_AT '_GE correct_JJ rules_NN2 '_GE of_IO the_AT language_NN1 in_II the_AT Chomskyan_JJ approach_NN1 ,_, and_CC that_DD1 of_IO most_DAT other_JJ sentence-grammarians_NN2 ,_, that_CST they_PPHS2 are_VBR based_VVN on_II the_AT presentation_NN1 of_IO '_GE example_NN1 '_GE and_CC '_GE counterexample_NN1 '_GE ._. 
After_II all_DB ,_, a_AT1 single_JJ (_( accepted_VVN )_) sentence_NN1 ,_, which_DDQ is_VBZ presented_VVN as_II a_AT1 counterexample_NN1 ,_, can_VM be_VBI enough_DD to_TO invalidate_VVI a_AT1 rule_NN1 of_IO the_AT categorial_JJ type_NN1 ._. 
In_II this_DD1 sense_NN1 ,_, the_AT '_GE rules_NN2 '_GE of_IO grammar_NN1 appear_VV0 to_TO be_VBI treated_VVN in_II the_AT same_DA way_NN1 as_CSA '_GE laws_NN2 '_GE in_II the_AT physical_JJ sciences_NN2 ._. 
This_DD1 restricts_VVZ the_AT applicability_NN1 of_IO such_DA rules_NN2 since_CS it_PPH1 renders_VVZ them_PPHO2 unavailable_JJ to_II any_DD linguist_NN1 interested_JJ in_II diachronic_JJ change_NN1 or_CC synchronic_JJ variation_NN1 in_II a_AT1 language_NN1 ._. 
It_PPH1 should_VM be_VBI emphasised_VVN that_CST this_DD1 is_VBZ an_AT1 extreme_JJ version_NN1 of_IO the_AT sentence-grammarian_NN1 's_GE view_NN1 and_CC one_PN1 which_DDQ is_VBZ found_VVN less_RGR frequently_RR ,_, in_II contemporary_JJ linguistics_NN1 ,_, than_CSN it_PPH1 was_VBDZ fifteen_MC years_NNT2 ago_RA ._. 
The_AT discourse_NN1 analyst_NN1 ,_, with_IW his_58 '_GE ordinary_JJ language_NN1 '_GE data_NN ,_, is_VBZ committed_VVN to_II quite_RG a_AT1 different_JJ view_NN1 of_IO the_AT rule-governed_JJ aspects_NN2 of_IO a_AT1 language_NN1 ._. 
Indeed_RR ,_, he_PPHS1 may_VM wish_VVI to_TO discuss_VVI ,_, not_XX '_GE rules_NN2 '_GE but_CCB regularities_NN2 ,_, simply_RR because_CS his_APPGE data_NN constantly_RR exemplifies_VVZ non-categorial_JJ phenomena_NN2 ._. 
The_AT regularities_NN2 which_DDQ the_AT analyst_NN1 describes_VVZ are_VBR based_VVN on_II the_AT frequency_NN1 with_IW which_DDQ a_AT1 particular_JJ linguistic_JJ feature_NN1 occurs_VVZ under_II certain_JJ conditions_NN2 in_II his_APPGE discourse_NN1 data_NN ._. 
If_CS the_AT frequency_NN1 of_IO occurrence_NN1 is_VBZ very_RG high_JJ ,_, then_RT the_AT phenomenon_NN1 described_VVN may_VM appear_VVI to_TO be_VBI categorial_JJ ._. 
As_CSA Givn_NP1 says_VVZ :_: what_DDQ is_VBZ the_AT communicative_JJ difference_NN1 between_II a_AT1 rule_NN1 of_IO 90%_NNU fidelity_NN1 and_CC one_MC1 of_IO 100%_NNU fidelity_NN1 ?_? 
In_II psychological_JJ terms_NN2 ,_, next_II21 to_II22 nothing_PN1 ._. 
In_II communication_NN1 ,_, a_AT1 system_NN1 with_IW go%_FO categorial_JJ fidelity_NN1 is_VBZ a_AT1 highly_RR efficient_JJ system_NN1 ._. 
(_( Givn_VV0 ,_, 1979a_FO :_: 28_MC )_) Yet_RR the_AT frequency_NN1 of_IO occurrence_NN1 need_VM not_XX be_VBI as_RG high_JJ as_CSA 90%_NNU to_TO qualify_VVI as_II a_AT1 regularity_NN1 ._. 
The_AT discourse_NN1 analyst_NN1 ,_, like_II the_AT experimental_JJ psychologist_NN1 ,_, is_VBZ mainly_RR interested_JJ in_II the_AT level_NN1 of_IO frequency_NN1 which_DDQ reaches_VVZ significance_NN1 in_II perceptual_JJ terms_NN2 ._. 
Thus_RR ,_, a_AT1 regularity_NN1 in_II discourse_NN1 is_VBZ a_AT1 linguistic_JJ feature_NN1 which_DDQ occurs_VVZ in_II a_AT1 definable_JJ environment_NN1 with_IW a_AT1 significant_JJ frequency_NN1 ._. 
In_II trying_VVG to_TO determine_VVI such_DA regularities_NN2 ,_, the_AT discourse_NN1 analyst_NN1 will_VM typically_RR adopt_VVI the_AT traditional_JJ methodology_NN1 of_IO descriptive_JJ linguistics_NN1 ._. 
He_PPHS1 will_VM attempt_VVI to_TO describe_VVI the_AT linguistic_JJ forms_NN2 which_DDQ occur_VV0 in_II his_APPGE data_NN ,_, relative_II21 to_II22 the_AT environments_NN2 in_II which_DDQ they_PPHS2 occur_VV0 ._. 
In_II this_DD1 sense_NN1 ,_, discourse_NN1 analysis_NN1 is_VBZ ,_, like_II descriptive_JJ linguistics_NN1 ,_, a_AT1 way_NN1 of_IO studying_VVG language_NN1 ._. 
It_PPH1 may_VM be_VBI regarded_VVN as_II a_AT1 set_NN1 of_IO techniques_NN2 ,_, rather_II21 than_II22 a_AT1 theoretically_RR predetermined_JJ system_NN1 for_IF the_AT writing_NN1 of_IO linguistic_JJ '_GE rules_NN2 '_GE ._. 
The_AT discourse_NN1 analyst_NN1 attempts_NN2 to_TO discover_VVI regularities_NN2 in_II his_APPGE data_NN and_CC to_TO describe_VVI them_PPHO2 ._. 
Product_NN1 versus_II process_NN1 The_AT regularities_NN2 which_DDQ the_AT discourse_NN1 analyst_NN1 describes_VVZ will_VM normally_RR be_VBI expressed_VVN in_II dynamic_JJ ,_, not_XX static_JJ ,_, terms_NN2 ._. 
Since_CS the_AT data_NN investigated_VVN is_VBZ the_AT result_NN1 of_IO '_GE ordinary_JJ language_NN1 behaviour_NN1 '_GE ,_, it_PPH1 is_VBZ likely_JJ to_TO contain_VVI evidence_NN1 of_IO the_AT '_GE behaviour_NN1 '_GE element_NN1 ._. 
That_REX21 is_REX22 ,_, unless_CS we_PPIS2 believe_VV0 that_CST language-users_NN2 present_VV0 each_PPX221 other_PPX222 with_IW prefabricated_JJ chunks_NN2 of_IO linguistic_JJ strings_NN2 (_( sentences_NN2 )_) ,_, after_CS the_AT fashion_NN1 of_IO Swift_NP1 's_GE professors_NN2 at_II the_AT grand_JJ academy_NN1 of_IO Lagado_NP1 (_( Gulliver_NP1 's_GE Travels_NN2 ,_, part_NN1 3_MC ,_, chapter_NN1 5_MC )_) ,_, then_RT we_PPIS2 must_VM assume_VVI that_CST the_AT data_NN we_PPIS2 investigate_VV0 is_VBZ the_AT result_NN1 of_IO active_JJ processes_NN2 ._. 
The_AT sentence-grammarian_NN1 does_VDZ not_XX in_RR21 general_RR22 take_VV0 account_NN1 of_IO this_DD1 ,_, since_CS his_APPGE data_NN is_VBZ not_XX connected_VVN to_II behaviour_NN1 ._. 
His_APPGE data_NN consists_VVZ of_IO a_AT1 set_NN1 of_IO objects_NN2 called_VVN '_GE the_AT well-formed_JJ sentences_NN2 of_IO a_AT1 language_NN1 '_GE ,_, which_DDQ can_VM exist_VVI independently_RR of_IO any_DD individual_JJ speaker_NN1 of_IO that_DD1 language_NN1 ._. 
We_PPIS2 shall_VM characterise_VVI such_DA a_AT1 view_NN1 as_CSA the_AT sentence-as-object_JJ view_NN1 ,_, and_CC note_VV0 that_CST such_DA sentenceobjects_NN2 have_VH0 no_AT producers_NN2 and_CC no_AT receivers_NN2 ._. 
Moreover_RR ,_, they_PPHS2 need_VM not_XX be_VBI considered_VVN in_II31 terms_II32 of_II33 function_NN1 ,_, as_CSA evidenced_VVN in_II this_DD1 statement_NN1 by_II Chomsky_NP1 (_( I968_FO :_: 62_MC )_) :_: If_CS we_PPIS2 hope_VV0 to_TO understand_VVI human_JJ language_NN1 and_CC the_AT psychological_JJ capacities_NN2 on_II which_DDQ it_PPH1 rests_VVZ ,_, we_PPIS2 must_VM first_MD ask_VVI what_DDQ it_PPH1 is_VBZ ,_, not_XX how_RRQ or_CC for_IF what_DDQ purposes_VVZ it_PPH1 is_VBZ used_VVN ._. 
A_AT1 less_RGR extreme_JJ ,_, but_CCB certainly_RR related_VVN ,_, view_NN1 of_IO natural_JJ language_NN1 sentences_NN2 can_VM also_RR be_VBI found_VVN elsewhere_RL in_II the_AT literature_NN1 which_DDQ relates_VVZ to_II discourse_NN1 analysis_NN1 ._. 
In_II this_DD1 view_NN1 ,_, there_EX are_VBR producers_NN2 and_CC receivers_NN2 of_IO sentences_NN2 ,_, or_CC extended_VVD texts_NN2 ,_, but_CCB the_AT analysis_NN1 concentrates_VVZ solely_RR on_II the_AT product_NN1 ,_, that_REX21 is_REX22 ,_, the_AT words-on-thepage_NN2 ._. 
Much_DA1 of_IO the_AT analytic_JJ work_NN1 undertaken_VVN in_II '_GE Textlinguistics_NP1 '_GE is_VBZ of_IO this_DD1 type_NN1 ._. 
Typical_JJ of_IO such_DA an_AT1 approach_NN1 is_VBZ the_AT '_GE cohesion_NN1 '_GE view_NN1 of_IO the_AT relationships_NN2 between_II sentences_NN2 in_II a_AT1 printed_JJ text_NN1 (_( e.g._REX the_AT approach_NN1 in_II Halliday_NP1 &amp;_CC Hasan_NP1 ,_, 1976_MC )_) ._. 
In_II this_DD1 view_NN1 ,_, cohesive_JJ ties_NN2 exist_VV0 between_II elements_NN2 in_II connected_JJ sentences_NN2 of_IO a_AT1 text_NN1 in_II such_DA a_AT1 way_NN1 that_CST one_MC1 word_NN1 or_CC phrase_NN1 is_VBZ linked_VVN to_II other_JJ words_NN2 or_CC phrases_NN2 ._. 
Thus_RR ,_, an_AT1 anaphoric_JJ element_NN1 such_II21 as_II22 a_AT1 pronoun_NN1 is_VBZ treated_VVN as_II a_AT1 word_NN1 which_DDQ substitutes_NN2 for_IF ,_, or_CC refers_VVZ back_RP to_II ,_, another_DD1 word_NN1 or_CC words_NN2 ._. 
Although_CS there_EX are_VBR claims_NN2 that_CST cohesive_JJ links_NN2 in_II texts_NN2 are_VBR used_VVN by_II text-producers_NN2 to_TO facilitate_VVI reading_NN1 or_CC comprehension_NN1 by_II text-receivers_NN2 (_( cf._VVI Rochester_NP1 &amp;_CC Martin_NP1 1977_MC ,_, 1979_MC ;_; Kllgren_NP1 ,_, 1979_MC )_) ,_, the_AT analysis_NN1 of_IO the_AT '_GE product_NN1 '_GE ,_, i.e._REX the_AT printed_JJ text_NN1 itself_PPX1 ,_, does_VDZ not_XX involve_VVI any_DD consideration_NN1 of_IO how_RRQ the_AT product_NN1 is_VBZ produced_VVN or_CC how_RRQ it_PPH1 is_VBZ received_VVN ._. 
We_PPIS2 shall_VM describe_VVI such_DA an_AT1 approach_NN1 as_II deriving_VVG from_II a_AT1 text-as-product_JJ view_NN1 ._. 
This_DD1 view_NN1 does_VDZ not_XX take_VVI account_NN1 of_IO those_DD2 principles_NN2 which_DDQ constrain_VV0 the_AT production_NN1 and_CC those_DD2 which_DDQ constrain_VV0 the_AT interpretation_NN1 of_IO texts_NN2 ._. 
In_II contrast_NN1 to_II these_DD2 two_MC broadly_RR defined_VVN approaches_NN2 ,_, the_AT view_NN1 taken_VVN in_II this_DD1 book_NN1 is_VBZ best_RRT characterised_VVN as_II a_AT1 discourse-as-process_JJ view_NN1 ._. 
The_AT distinction_NN1 between_II treating_VVG discourse_NN1 as_58 '_GE product_NN1 '_GE or_CC '_GE process_NN1 '_GE has_VHZ already_RR been_VBN made_VVN by_II Widdowson_NP1 (_( 1979b_FO :_: 71_MC )_) ._. 
We_PPIS2 shall_VM consider_VVI words_NN2 ,_, phrases_NN2 and_CC sentences_NN2 which_DDQ appear_VV0 in_II the_AT textual_JJ record_NN1 of_IO a_AT1 discourse_NN1 to_TO be_VBI evidence_NN1 of_IO an_AT1 attempt_NN1 by_II a_AT1 producer_NN1 (_( speaker_NN1 /_FO writer_NN1 )_) to_TO communicate_VVI his_APPGE message_NN1 to_II a_AT1 recipient_NN1 (_( hearer_NN1 /_FO reader_NN1 )_) ._. 
We_PPIS2 shall_VM be_VBI particularly_RR interested_JJ in_II discussing_VVG how_RRQ a_AT1 recipient_NN1 might_VM come_VVI to_TO comprehend_VVI the_AT producer_NN1 's_VBZ intended_VVN message_NN1 on_II a_AT1 particular_JJ occasion_NN1 ,_, and_CC how_RRQ the_AT requirements_NN2 of_IO the_AT particular_JJ recipient(s)_NN2 ,_, in_II definable_JJ circumstances_NN2 ,_, influence_VV0 the_AT organisation_NN1 of_IO the_AT producer_NN1 's_GE discourse_NN1 ._. 
This_DD1 is_VBZ clearly_RR an_AT1 approach_NN1 which_DDQ takes_VVZ the_AT communicative_JJ function_NN1 of_IO language_NN1 as_CSA its_APPGE primary_JJ area_NN1 of_IO investigation_NN1 and_CC consequently_RR seeks_VVZ to_TO describe_VVI linguistic_JJ form_NN1 ,_, not_XX as_II a_AT1 static_JJ object_NN1 ,_, but_CCB as_II a_AT1 dynamic_JJ means_NN of_IO expressing_VVG intended_JJ meaning_NN1 ._. 
There_EX are_VBR several_DA2 arguments_NN2 against_II the_AT static_JJ concept_NN1 of_IO language_NN1 to_TO be_VBI found_VVN in_II both_DB2 the_AT '_GE sentence-as-object_NN1 '_GE and_CC '_GE text-as-product_NN1 '_GE approaches_NN2 ._. 
For_REX21 example_REX22 ,_, Wittgenstein_NP1 (_( 1953_MC :_: 132_MC )_) warns_VVZ that_DD1 '_VBZ the_AT confusions_NN2 that_CST occupy_VV0 us_PPIO2 arise_VV0 when_RRQ language_NN1 is_VBZ like_II an_AT1 engine_NN1 idling_VVG ,_, not_XX when_RRQ it_PPH1 is_VBZ doing_VDG work_NN1 '_GE ._. 
In_II the_AT course_NN1 of_IO describing_VVG how_RRQ a_AT1 sentence-as-object_JJ approach_NN1 ,_, based_VVN exclusively_RR on_II syntactic_JJ descriptions_NN2 ,_, fails_VVZ to_TO account_VVI for_IF a_AT1 variety_NN1 of_IO sentential_JJ structures_NN2 ,_, Kuno_NP1 (_( 1976_MC )_) concludes_VVZ that_CST '_" it_PPH1 is_VBZ time_NNT1 to_TO re-examine_VVI every_AT1 major_JJ syntactic_JJ constraint_NN1 from_II a_AT1 functional_JJ point_NN1 of_IO view_NN1 '_GE ._. 
Similar_JJ conclusions_NN2 are_VBR expressed_VVN by_II Creider_NP1 (_( 1979_MC )_) ,_, Givon_NP1 (_( 1976_MC ,_, 1979b_FO )_) ,_, Rommetveit_NP1 (_( 1974_MC )_) and_CC Tyler_NP1 (_( 1978_MC )_) ._. 
In_II criticising_VVG the_AT text-as-product_JJ view_NN1 of_IO cohesion_NN1 in_II text_NN1 ,_, Morgan_NP1 (_( 1979_MC )_) argues_VVZ that_CST we_PPIS2 see_VV0 a_AT1 link_NN1 between_II a_AT1 particular_JJ pronoun_NN1 and_CC a_AT1 full_JJ noun_NN1 phrase_NN1 in_II a_AT1 text_NN1 because_CS we_PPIS2 assume_VV0 the_AT text_NN1 is_VBZ coherent_JJ and_CC not_XX because_CS the_AT pronoun_NN1 '_GE refers_VVZ back_NN1 '_GE to_II the_AT noun_NN1 phrase_NN1 ._. 
We_PPIS2 seek_VV0 to_TO identify_VVI the_AT writer_NN1 's_VBZ intended_VVN referent_NN1 for_IF a_AT1 pronoun_NN1 ,_, since_CS a_AT1 pronoun_NN1 can_VM ,_, in_II effect_NN1 ,_, be_VBI used_VVN to_TO refer_VVI to_II almost_RR anything_PN1 ._. 
That_REX21 is_REX22 ,_, what_DDQ the_AT textual_JJ record_NN1 means_NN is_VBZ determined_VVN by_II our_APPGE interpretation_NN1 of_IO what_DDQ the_AT producer_NN1 intended_VVD it_PPH1 to_TO mean_VVI ._. 
The_AT discourse_NN1 analyst_NN1 ,_, then_RT ,_, is_VBZ interested_JJ in_II the_AT function_NN1 or_CC purpose_NN1 of_IO a_AT1 piece_NN1 of_IO linguistic_JJ data_NN and_CC also_RR in_II how_RRQ that_DD1 data_NN is_VBZ processed_VVN ,_, both_RR by_II the_AT producer_NN1 and_CC by_II the_AT receiver_NN1 ._. 
It_PPH1 is_VBZ a_AT1 natural_JJ consequence_NN1 that_CST the_AT discourse_NN1 analyst_NN1 will_VM be_VBI interested_JJ in_II the_AT results_NN2 of_IO psycholinguistic_JJ processing_NN1 experiments_NN2 in_II a_AT1 way_NN1 which_DDQ is_VBZ not_XX typical_JJ of_IO the_AT sentence-grammarian_NN1 ._. 
It_PPH1 also_RR follows_VVZ that_CST the_AT work_NN1 of_IO those_DD2 sociolinguists_NN2 and_CC ethnographers_NN2 who_PNQS attempt_VV0 to_TO discuss_VVI language_NN1 in_II31 terms_II32 of_II33 user_NN1 's_GE purposes_NN2 will_VM also_RR be_VBI of_IO interest_NN1 ._. 
In_II the_AT course_NN1 of_IO this_DD1 book_NN1 ,_, we_PPIS2 shall_VM appeal_VVI to_II evidence_NN1 in_II the_AT psycholinguistic_JJ and_CC sociolinguistic_JJ literature_NN1 which_DDQ offers_VVZ insights_NN2 into_II the_AT way_NN1 in_II which_DDQ discourse_NN1 ,_, produced_VVN in_II describable_JJ contexts_NN2 for_IF recognisable_JJ purposes_NN2 ,_, is_VBZ processed_VVN and_CC comprehended_VVN ._. 
On_II '_GE context'_NN1 We_PPIS2 have_VH0 constantly_RR referred_VVN to_II the_AT '_GE environment_NN1 '_GE ,_, '_GE circumstances_NN2 '_GE or_CC context_NN1 in_II which_DDQ language_NN1 is_VBZ used_VVN ._. 
In_II Chapter_NN1 2_MC we_PPIS2 shall_VM explore_VVI the_AT problem_NN1 of_IO specifying_VVG the_AT relevant_JJ context_NN1 ._. 
Here_RL we_PPIS2 simply_RR remark_VV0 that_CST in_II recent_JJ years_NNT2 the_AT idea_NN1 that_CST a_AT1 linguistic_JJ string_NN1 (_( a_AT1 sentence_NN1 )_) can_VM be_VBI fully_RR analysed_VVN without_IW taking_NN1 '_GE context_NN1 '_GE into_II account_NN1 has_VHZ been_VBN seriously_RR questioned_VVN ._. 
If_CS the_AT sentence-grammarian_NN1 wishes_VVZ to_TO make_VVI claims_NN2 about_II the_AT '_GE acceptability_NN1 '_GE of_IO a_AT1 sentence_NN1 in_II determining_VVG whether_CSW the_AT strings_NN2 produced_VVN by_II his_APPGE grammar_NN1 are_VBR correct_JJ sentences_NN2 of_IO the_AT language_NN1 ,_, he_PPHS1 is_VBZ implicitly_RR appealing_VVG to_II contextual_JJ considerations_NN2 ._. 
After_II all_DB ,_, what_DDQ do_VD0 we_PPIS2 do_VDI when_CS we_PPIS2 are_VBR asked_VVN whether_CSW a_AT1 particular_JJ string_NN1 is_58 '_GE acceptable_JJ '_" ?_? 
Do_VD0 we_PPIS2 not_XX immediately_RR ,_, and_CC quite_RG naturally_RR ,_, set_VVD about_RP constructing_VVG some_DD circumstances_NN2 (_( i.e._REX a_AT1 '_GE context_NN1 '_GE )_) in_II which_DDQ the_AT sentence_NN1 could_VM be_VBI acceptably_RR used_VVN ?_? 
Any_DD analytic_JJ approach_NN1 in_II linguistics_NN1 which_DDQ involves_VVZ contextual_JJ considerations_NN2 ,_, necessarily_RR belongs_VVZ to_II that_DD1 area_NN1 of_IO language_NN1 study_NN1 called_VVN pragmatics_NN1 ._. 
'_" Doing_VDG discourse_NN1 analysis_NN1 '_GE certainly_RR involves_58 '_GE doing_VDG syntax_NN1 and_CC semantics_NN1 '_GE ,_, but_CCB it_PPH1 primarily_RR consists_VVZ of_IO '_" doing_VDG pragmatics_NN1 '_GE ._. 
When_CS the_AT principles_NN2 which_DDQ we_PPIS2 have_VH0 expounded_VVN in_II 1.3_MC are_VBR placed_VVN alongside_II Morris_NP1 's_GE definition_NN1 of_IO pragmatics_NN1 as_58 '_GE the_AT relations_NN2 of_IO signs_NN2 to_II interpreters_NN2 '_GE (_( 1938_MC :_: 6_MC )_) ,_, the_AT connection_NN1 becomes_VVZ quite_RG clear_JJ ._. 
In_II discourse_NN1 analysis_NN1 ,_, as_CSA in_II pragmatics_NN1 ,_, we_PPIS2 are_VBR concerned_JJ with_IW what_DDQ people_NN using_VVG language_NN1 are_VBR doing_VDG ,_, and_CC accounting_VVG for_IF the_AT linguistic_JJ features_NN2 in_II the_AT discourse_NN1 as_II the_AT means_NN employed_VVN in_II what_DDQ they_PPHS2 are_VBR doing_VDG ._. 
In_II summary_NN1 ,_, the_AT discourse_NN1 analyst_NN1 treats_VVZ his_APPGE data_NN as_II the_AT record_NN1 (_( text_NN1 )_) of_IO a_AT1 dynamic_JJ process_NN1 in_II which_DDQ language_NN1 was_VBDZ used_VVN as_II an_AT1 instrument_NN1 of_IO communication_NN1 in_II a_AT1 context_NN1 by_II a_AT1 speaker_NN1 /_FO writer_NN1 to_TO express_VVI meanings_NN2 and_CC achieve_VVI intentions_NN2 (_( discourse_NN1 )_) ._. 
Working_VVG from_II this_DD1 data_NN ,_, the_AT analyst_NN1 seeks_VVZ to_TO describe_VVI regularities_NN2 in_II the_AT linguistic_JJ realisations_NN2 used_VVN by_II people_NN to_TO communicate_VVI those_DD2 meanings_NN2 and_CC intentions._NNU 2_MC The_AT role_NN1 of_IO context_NN1 in_II interpretation_NN1 Pragmatics_NN1 and_CC discourse_NN1 context_NN1 In_II Chapter_NN1 I_ZZ1 ,_, we_PPIS2 emphasised_VVD that_CST the_AT discourse_NN1 analyst_NN1 necessarily_RR takes_VVZ a_AT1 pragmatic_JJ approach_NN1 to_II the_AT study_NN1 of_IO language_NN1 in_II use_NN1 ._. 
Such_DA an_AT1 approach_NN1 brings_VVZ into_II consideration_NN1 a_AT1 number_NN1 of_IO issues_NN2 which_DDQ do_VD0 not_XX generally_RR receive_VVI much_DA1 attention_NN1 in_II the_AT formal_JJ linguist_NN1 's_GE description_NN1 of_IO sentential_JJ syntax_NN1 and_CC semantics_NN1 ._. 
We_PPIS2 noted_VVD ,_, for_REX21 example_REX22 ,_, that_CST the_AT discourse_NN1 analyst_NN1 has_VHZ to_TO take_VVI account_NN1 of_IO the_AT context_NN1 in_II which_DDQ a_AT1 piece_NN1 of_IO discourse_NN1 occurs_VVZ ._. 
Some_DD of_IO the_AT most_RGT obvious_JJ linguistic_JJ elements_NN2 which_DDQ require_VV0 contextual_JJ information_NN1 for_IF their_APPGE interpretation_NN1 are_VBR the_AT deictic_JJ forms_NN2 such_II21 as_II22 here_RL ,_, now_RT ,_, 1_MC1 ,_, you_PPY ,_, this_DD1 and_CC that_DD1 ._. 
In_BCL21 order_BCL22 to_TO interpret_VVI these_DD2 elements_NN2 in_II a_AT1 piece_NN1 of_IO discourse_NN1 ,_, it_PPH1 is_VBZ necessary_JJ to_TO know_VVI (_( at_RR21 least_RR22 )_) who_PNQS the_AT speaker_NN1 and_CC hearer_NN1 are_VBR ,_, and_CC the_AT time_NNT1 and_CC place_NN1 of_IO the_AT production_NN1 of_IO the_AT discourse_NN1 ._. 
In_II this_DD1 chapter_NN1 we_PPIS2 shall_VM discuss_VVI these_DD2 and_CC other_JJ aspects_NN2 of_IO contextual_JJ description_NN1 which_DDQ are_VBR required_VVN in_II the_AT analysis_NN1 of_IO discourse_NN1 ._. 
There_EX are_VBR ,_, however_RR ,_, other_JJ ways_NN2 in_II which_DDQ the_AT discourse_NN1 analyst_NN1 's_GE approach_NN1 to_II linguistic_JJ data_NN differs_VVZ from_II that_DD1 of_IO the_AT formal_JJ linguist_NN1 and_CC leads_VVZ to_II a_AT1 specialised_JJ use_NN1 of_IO certain_JJ terms_NN2 ._. 
Because_CS the_AT analyst_NN1 is_VBZ investigating_VVG the_AT use_NN1 of_IO language_NN1 in_II context_NN1 by_II a_AT1 speaker_NN1 /_FO writer_NN1 ,_, he_PPHS1 is_VBZ more_RGR concerned_JJ with_IW the_AT relationship_NN1 between_II the_AT speaker_NN1 and_CC the_AT utterance_NN1 ,_, on_II the_AT particular_JJ occasion_NN1 of_IO use_NN1 ,_, than_CSN with_IW the_AT potential_JJ relationship_NN1 of_IO one_MC1 sentence_NN1 to_II another_DD1 ,_, regardless_RR of_IO their_APPGE use_NN1 ._. 
That_REX21 is_REX22 ,_, in_II using_VVG terms_NN2 such_II21 as_II22 reference_NN1 ,_, presupposition_NN1 ,_, implicature_NN1 and_CC inference_NN1 ,_, the_AT discourse_NN1 analyst_NN1 is_VBZ describing_VVG what_DDQ speakers_NN2 and_CC hearers_NN2 are_VBR doing_VDG ,_, and_CC not_XX the_AT relationship_NN1 which_DDQ exists_VVZ between_II one_MC1 sentence_NN1 or_CC proposition_NN1 and_CC another_DD1 ._. 
Reference_NN1 In_II presenting_VVG the_AT traditional_JJ semantic_JJ view_NN1 of_IO reference_NN1 ,_, Lyons_NP1 (_( I968_FO :_: 404_MC )_) says_VVZ that_DD1 '_VBZ the_AT relationship_NN1 which_DDQ holds_VVZ between_II words_NN2 and_CC things_NN2 is_VBZ the_AT relationship_NN1 of_IO reference_NN1 :_: words_NN2 refer_VV0 to_II things_NN2 '_GE ._. 
This_DD1 traditional_JJ view_NN1 continues_VVZ to_TO be_VBI expressed_VVN in_II those_DD2 linguistic_JJ studies_NN2 (_( e.g._REX lexical_JJ semantics_NN1 )_) which_DDQ describe_VV0 the_AT relationship_NN1 between_II a_AT1 language_NN1 and_CC the_AT world_NN1 ,_, in_II the_AT absence_NN1 of_IO language-users_NN2 ._. 
Yet_RR ,_, Lyons_NP1 ,_, in_II a_AT1 more_RGR recent_JJ statement_NN1 on_II the_AT nature_NN1 of_IO reference_NN1 ,_, makes_VVZ the_AT following_JJ point_NN1 :_: '_" it_PPH1 is_VBZ the_AT speaker_NN1 who_PNQS refers_VVZ (_( by_II using_VVG some_DD appropriate_JJ expression_NN1 )_) :_: he_PPHS1 invests_VVZ the_AT expression_NN1 with_IW reference_NN1 by_II the_AT act_NN1 of_IO referring_VVG '_GE (_( 1977_MC :_: 177_MC )_) ._. 
It_PPH1 is_VBZ exactly_RR this_DD1 latter_DA view_NN1 of_IO the_AT nature_NN1 of_IO reference_NN1 which_DDQ the_AT discourse_NN1 analyst_NN1 has_VHZ to_TO appeal_VVI to_II ._. 
There_EX is_VBZ support_NN1 for_IF such_DA a_AT1 pragmatic_JJ concept_NN1 of_IO reference_NN1 in_II Strawson_NP1 's_GE (_( 1950_MC )_) claim_VV0 that_DD1 '_VBZ 'referring_NN1 '_GE is_VBZ not_XX something_PN1 an_AT1 expression_NN1 does_VDZ ;_; it_PPH1 is_VBZ something_PN1 that_CST someone_PN1 can_VM use_VVI an_AT1 expression_NN1 to_TO do'_VVI ;_; and_CC in_II Searle_NP1 's_GE view_NN1 that_DD1 '_VBZ in_II the_AT sense_NN1 in_II which_DDQ speakers_NN2 refer_VV0 ,_, expressions_NN2 do_VD0 not_XX refer_VVI any_DD more_DAR than_CSN they_PPHS2 make_VV0 promises_NN2 or_CC give_VV0 orders_NN2 '_GE (_( 1979_MC :_: 155_MC )_) ._. 
Thus_RR ,_, in_II discourse_NN1 analysis_NN1 ,_, reference_NN1 is_VBZ treated_VVN as_II an_AT1 action_NN1 on_II41 the_II42 part_II43 of_II44 the_AT speaker_NN1 /_FO writer_NN1 ._. 
In_II the_AT following_JJ conversational_JJ fragment_NN1 ,_, we_PPIS2 shall_VM say_VVI ,_, for_REX21 example_REX22 ,_, that_DD1 speaker_NN1 A_ZZ1 uses_VVZ the_AT expressions_NN2 my_APPGE uncle_NN1 and_CC he_PPHS1 to_TO refer_VVI to_II one_MC1 individual_NN1 and_CC my_APPGE mother_NN1 's_GE sister_NN1 and_CC she_PPHS1 to_TO refer_VVI to_II another_DD1 ._. 
We_PPIS2 will_VM not_XX ,_, for_REX21 example_REX22 ,_, say_VV0 that_CST he_PPHS1 '_" refers_VVZ to_II '_" my_APPGE uncle_NN1 ._. 
(_( I_ZZ1 )_) A_ZZ1 :_: my_APPGE uncle_NN1 's_VBZ coming_VVG home_RL from_II Canada_NP1 on_II Sunday_NP1 +_FO he_PPHS1 's_VBZ due_JJ in_II +_FO B_ZZ1 :_: how_RGQ long_RR has_VHZ he_PPHS1 been_VBN away_RL for_IF or_CC has_VHZ he_PPHS1 just_RR been_VBN away_RL ?_? 
A_ZZ1 :_: Oh_UH no_UH they_PPHS2 lived_VVD in_II Canada_NP1 eh_UH he_PPHS1 was_VBDZ married_JJ to_II my_APPGE mother_NN1 's_GE sister_NN1 +_FO +_FO well_RR she_PPHS1 's_VHZ been_VBN dead_JJ for_IF a_AT1 number_NN1 of_IO years_NNT2 now_RT +_FO The_AT complex_JJ nature_NN1 of_IO discourse_NN1 reference_NN1 will_VM be_VBI investigated_VVN in_II greater_JJR detail_NN1 in_II Chapters_NN2 5_MC and_CC 6_MC ._. 
Presupposition_NN1 In_II the_AT preceding_JJ conversational_JJ fragment_NN1 (_( I_ZZ1 )_) ,_, we_PPIS2 shall_VM also_RR say_VVI that_DD1 speaker_NN1 A_ZZ1 treats_VVZ the_AT information_NN1 that_CST she_PPHS1 has_VHZ an_AT1 uncle_NN1 as_CSA presupposed_VVN and_CC speaker_NN1 B_ZZ1 ,_, in_II her_APPGE question_NN1 ,_, indicates_VVZ that_CST she_PPHS1 has_VHZ accepted_VVN this_DD1 presupposition_NN1 ._. 
We_PPIS2 shall_VM take_VVI the_AT view_NN1 that_CST the_AT notion_NN1 of_IO presupposition_NN1 required_VVN in_II discourse_NN1 analysis_NN1 is_VBZ pragmatic_JJ presupposition_NN1 ,_, that_REX21 is_REX22 ,_, '_GE defined_JJ in_II31 terms_II32 of_II33 assumptions_NN2 the_AT speaker_NN1 makes_VVZ about_II what_DDQ the_AT hearer_NN1 is_VBZ likely_JJ to_TO accept_VVI without_IW challenge_NN1 '_GE (_( Givn_NP1 ,_, 1979a_FO :_: 50_MC )_) ._. 
The_AT notion_NN1 of_IO assumed_JJ '_GE common_JJ ground_NN1 '_GE is_VBZ also_RR involved_JJ in_II such_DA a_AT1 characterisation_NN1 of_IO presupposition_NN1 and_CC can_VM be_VBI found_VVN in_II this_DD1 definition_NN1 by_II Stalnaker_NP1 (_( 1978_MC :_: 321_MC )_) :_: presuppositions_NN2 are_VBR what_DDQ is_VBZ taken_VVN by_II the_AT speaker_NN1 to_TO be_VBI the_AT common_JJ ground_NN1 of_IO the_AT participants_NN2 in_II the_AT conversation_NN1 ._. 
Notice_VV0 that_CST ,_, in_II both_DB2 these_DD2 quotations_NN2 ,_, the_AT indicated_JJ source_NN1 of_IO presuppositions_NN2 is_VBZ the_AT speaker_NN1 ._. 
Consequently_RR ,_, we_PPIS2 shall_VM ,_, as_CSA with_IW reference_NN1 ,_, avoid_VV0 attributing_VVG presuppositions_NN2 to_II sentences_NN2 or_CC propositions_NN2 ._. 
Thus_RR ,_, we_PPIS2 can_VM see_VVI little_DA1 practical_JJ use_NN1 ,_, in_II the_AT analysis_NN1 of_IO discourse_NN1 ,_, for_IF the_AT notion_NN1 of_IO logical_JJ presupposition_NN1 which_DDQ Keenan_NP1 (_( 1971_MC :_: 45_MC )_) describes_VVZ in_II the_AT following_JJ way_NN1 :_: A_AT1 sentence_NN1 S_ZZ1 logically_RR presupposes_VVZ a_AT1 sentence_NN1 S'_NP1 just_RR in_CS21 case_CS22 S_ZZ1 logically_RR implies_VVZ S'_NP1 and_CC the_AT negation_NN1 of_IO S_ZZ1 ,_, S_ZZ1 ,_, also_RR logically_RR implies_VVZ S'_NP1 ._. 
If_CS we_PPIS2 take_VV0 the_AT first_MD sentence_NN1 of_IO extract_NN1 (_( I_ZZ1 )_) as_CSA S_ZZ1 ,_, and_CC present_VV0 it_PPH1 below_RL as_CSA (_( 2a_FO )_) ,_, we_PPIS2 can_VM also_RR present_VVI the_AT negation_NN1 of_IO S_ZZ1 ,_, as_CSA (_( 2b_FO )_) ,_, and_CC the_AT logical_JJ presupposition_NN1 ,_, S'_NP1 ,_, as_CSA (_( 2c_FO )_) ._. 
(_( 2_MC )_) a_AT1 ._. 
My_APPGE uncle_NN1 is_VBZ coming_VVG home_RL from_II Canada._NP1 b_ZZ1 ._. 
My_APPGE uncle_NN1 is_VBZ n't_XX coming_VVG home_RL from_II Canada._NP1 c_ZZ1 ._. 
I_PPIS1 have_VH0 an_AT1 uncle_NN1 ._. 
Following_VVG Keenan_NP1 's_GE definition_NN1 ,_, we_PPIS2 can_VM say_VVI that_CST (_( 2a_FO )_) logically_RR presupposes_VVZ (_( 2c_FO )_) because_II21 of_II22 constancy_NN1 under_II negation_NN1 ._. 
However_RR ,_, it_PPH1 seems_VVZ rather_RG unnecessary_JJ to_TO introduce_VVI the_AT negative_JJ sentence_NN1 (_( 2b_FO )_) into_II a_AT1 consideration_NN1 of_IO the_AT relationship_NN1 between_II (_( 2a_FO )_) and_CC (_( 2c_FO )_) which_DDQ arises_VVZ in_II the_AT conversation_NN1 presented_VVD earlier_RRR in_II (_( I_ZZ1 )_) ._. 
Though_CS it_PPH1 may_VM not_XX be_VBI common_JJ knowledge_NN1 that_CST the_AT speaker_NN1 has_VHZ an_AT1 uncle_NN1 ,_, it_PPH1 is_VBZ what_DDQ Grice_NP1 (_( 1981_MC :_: 190_MC )_) terms_NN2 '_GE noncontroversial_JJ '_GE information_NN1 ._. 
Moreover_RR ,_, since_CS the_AT speaker_NN1 chose_VVD to_TO say_VVI my_APPGE uncle_NN1 rather_CS21 than_CS22 I_PPIS1 have_VH0 an_AT1 uncle_NN1 and_CC he_PPHS1 ..._... ,_, we_PPIS2 must_VM assume_VVI she_PPHS1 did_VDD n't_XX feel_VVI the_AT need_NN1 to_TO assert_VVI the_AT information_NN1 ._. 
What_DDQ she_PPHS1 appears_VVZ to_TO be_VBI asserting_VVG is_VBZ that_CST this_DD1 person_NN1 is_VBZ coming_VVG home_RL from_II Canada_NP1 ._. 
Given_VVN this_DD1 assertion_NN1 ,_, the_AT idea_NN1 that_CST we_PPIS2 should_VM consider_VVI the_AT denial_NN1 of_IO this_DD1 assertion_NN1 in_BCL21 order_BCL22 to_TO find_VVI out_RP whether_CSW there_EX is_VBZ a_AT1 presupposition_NN1 in_II what_DDQ the_AT speaker_NN1 has_VHZ not_XX asserted_VVN seems_VVZ particularly_RR counterintuitive_JJ ._. 
The_AT introduction_NN1 of_IO the_AT negative_JJ sentence_NN1 (_( 2b_FO )_) into_II a_AT1 consideration_NN1 of_IO (_( 2a_FO )_) creates_VVZ an_AT1 additional_JJ problem_NN1 ._. 
For_REX21 example_REX22 ,_, it_PPH1 has_VHZ been_VBN suggested_VVN (_( cf._VV0 Kempson_NP1 ,_, 1975_MC )_) that_CST a_AT1 sentence_NN1 such_II21 as_II22 (_( 2d_NNU )_) is_VBZ a_AT1 perfectly_RR reasonable_JJ sentence_NN1 of_IO English_NN1 and_CC undermines_VVZ the_AT argument_NN1 for_IF logical_JJ presupposition_NN1 ,_, as_CSA it_PPH1 is_VBZ defined_VVN above_RL (_( 2d_NNU )_) My_APPGE uncle_NN1 is_VBZ n't_XX coming_VVG home_RL from_II Canada_NP1 because_CS I_PPIS1 do_VD0 n't_XX have_VHI an_AT1 uncle_NN1 ._. 
Sentences_NN2 like_II (_( 2d_NNU )_) always_RR seem_VV0 typical_JJ of_IO utterances_NN2 made_VVN by_II a_AT1 speaker_NN1 to_TO deny_VVI another_DD1 speaker_NN1 's_GE presupposition_NN1 in_II a_AT1 rather_RG aggressive_JJ way_NN1 ._. 
Yet_RR the_AT circumstances_NN2 in_II which_DDQ (_( 2d_NNU )_) might_VM be_VBI uttered_VVN are_VBR likely_JJ to_TO be_VBI quite_RG different_JJ from_II those_DD2 in_II which_DDQ the_AT first_MD sentence_NN1 of_IO extract_NN1 (_( I_ZZ1 )_) was_VBDZ uttered_VVN ._. 
The_AT speakers_NN2 ,_, we_PPIS2 may_VM suggest_VVI ,_, would_VM have_VHI different_JJ presuppositions_NN2 ,_, in_II the_AT two_MC situations_NN2 ._. 
If_CS we_PPIS2 rely_VV0 on_II a_AT1 notion_NN1 of_IO speaker_NN1 ,_, or_CC pragmatic_JJ ,_, presupposition_NN1 ,_, we_PPIS2 can_VM simply_RR treat_VVI (_( 2c_FO )_) as_II a_AT1 presupposition_NN1 of_IO the_AT speaker_NN1 in_II uttering_VVG (_( 2a_FO )_) ._. 
Sentences_NN2 (_( 2b_FO )_) and_CC (_( 2d_NNU )_) do_VD0 not_XX come_VVI into_II consideration_NN1 at_RR21 all_RR22 ._. 
In_II31 support_II32 of_II33 a_AT1 view_NN1 that_CST hearers_NN2 behave_VV0 as_CS21 if_CS22 speakers_NN2 '_GE presuppositions_NN2 are_VBR to_TO be_VBI accepted_VVN ,_, there_EX is_VBZ the_AT rather_RG disturbing_JJ evidence_NN1 from_II Loftus_NP1 '_GE study_NN1 (_( 1975_MC )_) of_IO answers_NN2 to_II leading_JJ questions_NN2 ._. 
After_II watching_VVG a_AT1 film_NN1 of_IO a_AT1 car_NN1 accident_NN1 some_DD subjects_NN2 were_VBDR asked_VVN the_AT two_MC questions_NN2 in_II (_( 3_MC )_) ._. 
(_( 3_MC )_) a_AT1 ._. 
How_RGQ fast_RR was_VBDZ car_NN1 A_ZZ1 going_VVG when_CS it_PPH1 turned_VVD right_RR ?_? b_ZZ1 ._. 
Did_VDD you_PPY see_VVI a_AT1 stop_NN1 sign_NN1 ?_? 
We_PPIS2 can_VM note_VVI that_CST one_MC1 of_IO the_AT speaker-presuppositions_NN2 in_II asking_VVG (_( 3a_FO )_) is_VBZ that_DD1 car_NN1 A_ZZ1 turned_VVD right_RR ._. 
A_AT1 number_NN1 (_( 35%_NNU )_) answered_VVD yes_UH to_TO question_VVI (_( 3b_FO )_) ._. 
Another_DD1 group_NN1 of_IO subjects_NN2 were_VBDR asked_VVN the_AT questions_NN2 in_II (_( 4_MC )_) ._. 
(_( 4_MC )_) a_AT1 ._. 
How_RGQ fast_RR was_VBDZ car_NN1 A_ZZ1 going_VVG when_CS it_PPH1 ran_VVD the_AT stop_NN1 sign_NN1 ?_? b_ZZ1 ._. 
Did_VDD you_PPY see_VVI a_AT1 stop_NN1 sign_NN1 ?_? 
One_MC1 of_IO the_AT speaker-presuppositions_NN2 in_II asking_VVG (_( 4a_FO )_) is_VBZ that_DD1 car_NN1 A_ZZ1 ran_VVD the_AT stop_NN1 sign_NN1 ._. 
In_II this_DD1 situation_NN1 ,_, a_AT1 significantly_RR larger_JJR group_NN1 (_( 53%_NNU )_) answered_VVD yes_UH to_TO question_VVI (_( 4b_FO )_) ._. 
It_PPH1 is_VBZ worth_II noting_VVG that_CST a_AT1 number_NN1 of_IO subjects_NN2 did_VDD not_XX answer_VVI the_AT b_ZZ1 question_NN1 in_II31 terms_II32 of_II33 truth_NN1 or_CC falsehood_NN1 of_IO fact_NN1 ,_, but_CCB according_II21 to_II22 what_DDQ the_AT speaker_NN1 ,_, in_II asking_VVG the_AT preceding_JJ question_NN1 ,_, had_VHD appeared_VVN to_TO presuppose_VVI ._. 
(_( For_IF a_AT1 more_RGR detailed_JJ discussion_NN1 of_IO this_DD1 issue_NN1 ,_, see_VV0 Loftus_NP1 ,_, 1975_MC and_CC Loftus_NP1 &amp;_CC Zanni_NP1 ,_, 1975_MC ._. )_) 
We_PPIS2 shall_VM reconsider_VVI the_AT notion_NN1 of_IO presupposition_NN1 in_II section_NN1 3.3.2_MC ,_, but_CCB generally_RR avoid_VV0 the_AT complex_JJ arguments_NN2 which_DDQ revolve_VV0 around_RP the_AT presuppositions_NN2 of_IO sentences_NN2 and_CC propositions_NN2 ._. 
(_( See_VV0 the_AT contributions_NN2 and_CC bibliography_NN1 in_II Oh_UH &amp;_CC Dineen_NP1 (_( eds._NN2 )_) 1979_MC ._. )_) 
Implicatures_NN2 The_AT term_NN1 '_GE implicature_NN1 '_GE is_VBZ used_VVN by_II Grice_NP1 (_( 1975_MC )_) to_TO account_VVI for_IF what_DDQ a_AT1 speaker_NN1 can_VM imply_VVI ,_, suggest_VVI ,_, or_CC mean_VVI ,_, as_CSA distinct_JJ from_II what_DDQ the_AT speaker_NN1 literally_RR says_VVZ ._. 
There_EX are_VBR conventional_JJ implicatures_NN2 which_DDQ are_VBR ,_, according_II21 to_II22 Grice_NP1 ,_, determined_VVN by_II '_GE the_AT conventional_JJ meaning_NN1 of_IO the_AT words_NN2 used_JJ '_GE (_( 1975_MC :_: 44_MC )_) ._. 
In_II the_AT following_JJ example_NN1 (_( 5_MC )_) ,_, the_AT speaker_NN1 does_VDZ not_XX directly_RR assert_VVI that_CST one_MC1 property_NN1 (_( being_VBG brave_JJ )_) follows_VVZ from_II another_DD1 property_NN1 (_( being_VBG an_AT1 Englishman_NN1 )_) ,_, but_CCB the_AT form_NN1 of_IO expression_NN1 used_VVD conventionally_RR implicates_VVZ that_CST such_DA a_AT1 relation_NN1 does_VDZ hold._NNU (_( 5_MC )_) He_PPHS1 is_VBZ an_AT1 Englishman_NN1 ,_, he_PPHS1 is_VBZ ,_, therefore_RR ,_, brave_JJ ._. 
If_CS it_PPH1 should_VM turn_VVI out_RP that_CST the_AT individual_NN1 in_II question_NN1 is_VBZ an_AT1 Englishman_NN1 ,_, and_CC not_XX brave_JJ ,_, then_RT the_AT implicature_NN1 is_VBZ mistaken_VVN ,_, but_CCB the_AT utterance_NN1 ,_, Grice_NP1 suggests_VVZ ,_, need_VV0 not_XX be_VBI false_JJ ._. 
For_IF a_AT1 fuller_JJR discussion_NN1 of_IO conventional_JJ implicature_NN1 ,_, see_VV0 Karttunen_JJ &amp;_CC Peters_VVZ (_( 1979_MC )_) ._. 
Of_IO much_DA1 greater_JJR interest_NN1 to_II the_AT discourse_NN1 analyst_NN1 is_VBZ the_AT notion_NN1 of_IO conversational_JJ implicature_NN1 which_DDQ is_VBZ derived_VVN from_II a_AT1 general_JJ principle_NN1 of_IO conversation_NN1 plus_II a_AT1 number_NN1 of_IO maxims_NN2 which_DDQ speakers_NN2 will_VM normally_RR obey_VVI ._. 
The_AT general_JJ principle_NN1 is_VBZ called_VVN the_AT Cooperative_JJ Principle_NN1 which_DDQ Grice_NP1 (_( 1975_MC :_: 45_MC )_) presents_VVZ in_II the_AT following_JJ terms_NN2 :_: Make_VV0 your_APPGE conversational_JJ contribution_NN1 such_II21 as_II22 is_VBZ required_VVN ,_, at_II the_AT stage_NN1 at_II which_DDQ it_PPH1 occurs_VVZ ,_, by_II the_AT accepted_JJ purpose_NN1 or_CC direction_NN1 of_IO the_AT talk_NN1 exchange_NN1 in_II which_DDQ you_PPY are_VBR engaged_VVN ._. 
The_AT conversational_JJ conventions_NN2 ,_, or_CC maxims_NN2 ,_, which_DDQ support_VV0 this_DD1 principle_NN1 are_VBR as_CSA follows_VVZ :_: Quantity_NN1 :_: Make_VV0 your_APPGE contribution_NN1 as_RG informative_JJ as_CSA is_VBZ required_VVN (_( for_IF the_AT current_JJ purposes_NN2 of_IO the_AT exchange_NN1 )_) ._. 
Do_VD0 not_XX make_VVI your_APPGE contribution_NN1 more_RGR informative_JJ than_CSN is_VBZ required_VVN ._. 
Quality_NN1 :_: Do_VD0 not_XX say_VVI what_DDQ you_PPY believe_VV0 to_TO be_VBI false_JJ ._. 
Do_VD0 not_XX say_VVI that_DD1 for_IF which_DDQ you_PPY lack_VV0 adequate_JJ evidence_NN1 ._. 
Relation_NN1 :_: Be_VBI relevant_JJ ._. 
Manner_NN1 :_: Be_VBI perspicuous_JJ ._. 
Avoid_VV0 obscurity_NN1 of_IO expression_NN1 ._. 
Avoid_VV0 ambiguity_NN1 ._. 
Be_VB0 brief_JJ (_( avoid_VV0 unnecessary_JJ prolixity_NN1 )_) ._. 
Be_VB0 orderly_JJ ._. 
Grice_NN1 does_VDZ not_XX suggest_VVI that_CST this_DD1 is_VBZ an_AT1 exhaustive_JJ list_NN1 he_PPHS1 notes_VVZ that_CST a_AT1 maxim_NN1 such_II21 as_II22 Be_VBI polite_JJ is_VBZ also_RR normally_RR observed_VVN nor_CC that_DD1 equal_JJ weight_NN1 should_VM be_VBI attached_VVN to_II each_DD1 of_IO the_AT stated_JJ maxims_NN2 ._. 
(_( The_AT maxim_NN1 of_IO manner_NN1 ,_, for_REX21 example_REX22 ,_, does_VDZ not_XX obviously_RR apply_VVI to_II primarily_RR interactional_JJ conversation_NN1 ._. )_) 
We_PPIS2 might_VM observe_VVI that_CST the_AT instruction_NN1 Be_VBI relevant_JJ seems_VVZ to_TO cover_VVI all_DB the_AT other_JJ instructions_NN2 ._. 
However_RR ,_, by_II providing_VVG a_AT1 description_NN1 of_IO the_AT norms_NN2 speakers_NN2 operate_VV0 with_IW in_II conversation_NN1 ,_, Grice_NP1 makes_VVZ it_PPH1 possible_JJ to_TO describe_VVI what_DDQ types_NN2 of_IO meaning_VVG a_AT1 speaker_NN1 can_VM convey_VVI by_II '_GE flouting_NN1 '_GE one_MC1 of_IO these_DD2 maxims_NN2 ._. 
This_DD1 flouting_NN1 of_IO a_AT1 maxim_NN1 results_NN2 in_II the_AT speaker_NN1 conveying_VVG ,_, in_II31 addition_II32 to_II33 the_AT literal_JJ meaning_NN1 of_IO his_APPGE utterance_NN1 ,_, an_AT1 additional_JJ meaning_NN1 ,_, which_DDQ is_VBZ a_AT1 conversational_JJ implicature_NN1 ._. 
As_II a_AT1 brief_JJ example_NN1 ,_, we_PPIS2 can_VM consider_VVI the_AT following_JJ exchange_NN1 :_: (_( 6_MC )_) A_ZZ1 :_: I_PPIS1 am_VBM out_II21 of_II22 petrol_NN1 ._. 
B_ZZ1 :_: There_EX is_VBZ a_AT1 garage_NN1 round_II the_AT corner_NN1 ._. 
In_II this_DD1 exchange_NN1 ,_, Grice_NP1 (_( 1975_MC :_: 51_MC )_) suggests_VVZ that_CST B_ZZ1 would_VM be_VBI infringing_VVG the_AT instruction_NN1 Be_VBI relevant_JJ if_CS he_PPHS1 was_VBDZ gratuitously_RR stating_VVG a_AT1 fact_NN1 about_II the_AT world_NN1 via_II the_AT literal_JJ meaning_NN1 of_IO his_APPGE utterance_NN1 ._. 
The_AT implicature_NN1 ,_, derived_VVN from_II the_AT assumption_NN1 that_CST speaker_NN1 B_ZZ1 is_VBZ adhering_VVG to_II the_AT Cooperative_JJ Principle_NN1 ,_, is_VBZ that_DD1 the_AT garage_NN1 is_VBZ not_XX only_RR round_II the_AT corner_NN1 ,_, but_CCB also_RR will_VM be_VBI open_JJ and_CC selling_VVG petrol_NN1 ._. 
We_PPIS2 might_VM also_RR note_VVI that_DD1 ,_, in_BCL21 order_BCL22 to_TO arrive_VVI at_II the_AT implicature_NN1 ,_, we_PPIS2 have_VH0 to_TO know_VVI certain_JJ facts_NN2 about_II the_AT world_NN1 ,_, that_CST garages_NN2 sell_VV0 petrol_NN1 ,_, and_CC that_CST round_II the_AT corner_NN1 is_VBZ not_XX a_AT1 great_JJ distance_NN1 away_RL ._. 
We_PPIS2 also_RR have_VH0 to_TO interpret_VVI A_ZZ221 's_ZZ222 remark_VV0 not_XX only_RR as_II a_AT1 description_NN1 of_IO a_AT1 particular_JJ state_NN1 of_IO affairs_NN2 ,_, but_CCB as_II a_AT1 request_NN1 for_IF help_NN1 ,_, for_REX21 instance_REX22 ._. 
Once_CS the_AT analysis_NN1 of_IO intended_JJ meaning_NN1 goes_VVZ beyond_II the_AT literal_JJ meaning_NN1 of_IO the_AT '_GE sentences-on-the-page_NN2 '_GE ,_, a_AT1 vast_JJ number_NN1 of_IO related_JJ issues_NN2 have_VH0 to_TO be_VBI considered_VVN ._. 
We_PPIS2 shall_VM investigate_VVI some_DD of_IO these_DD2 issues_NN2 in_II the_AT course_NN1 of_IO this_DD1 book_NN1 ,_, particularly_RR in_II Chapters_NN2 6_MC and_CC 7_MC ._. 
As_II a_AT1 brief_JJ account_NN1 of_IO how_RRQ the_AT term_NN1 '_GE implicature_NN1 '_GE is_VBZ used_VVN in_II discourse_NN1 analysis_NN1 ,_, we_PPIS2 have_VH0 summarised_VVN the_AT important_JJ points_NN2 in_II Grice_NP1 's_GE proposal_NN1 ._. 
We_PPIS2 would_VM like_VVI to_TO emphasise_VVI the_AT fact_NN1 that_CST implicatures_NN2 are_VBR pragmatic_JJ aspects_NN2 of_IO meaning_NN1 and_CC have_VH0 certain_JJ identifiable_JJ characteristics_NN2 ._. 
They_PPHS2 are_VBR partially_RR derived_VVN from_II the_AT conventional_JJ or_CC literal_JJ meaning_NN1 of_IO an_AT1 utterance_NN1 ,_, produced_VVN in_II a_AT1 specific_JJ context_NN1 which_DDQ is_VBZ shared_VVN by_II the_AT speaker_NN1 and_CC the_AT hearer_NN1 ,_, and_CC depend_VV0 on_II a_AT1 recognition_NN1 by_II the_AT speaker_NN1 and_CC the_AT hearer_NN1 of_IO the_AT Cooperative_JJ Principle_NN1 and_CC its_APPGE maxims_NN2 ._. 
For_IF the_AT analyst_NN1 ,_, as_II31 well_II32 as_II33 the_AT hearer_NN1 ,_, conversational_JJ implicatures_NN2 must_VM be_VBI treated_VVN as_CSA inherently_RR indeterminate_JJ since_CS they_PPHS2 derive_VV0 from_II a_AT1 supposition_NN1 that_CST the_AT speaker_NN1 has_VHZ the_AT intention_NN1 of_IO conveying_VVG meaning_NN1 and_CC of_IO obeying_VVG the_AT Cooperative_JJ Principle_NN1 ._. 
Since_CS the_AT analyst_NN1 has_VHZ only_RR limited_VVN access_NN1 to_II what_DDQ a_AT1 speaker_NN1 intended_VVD ,_, or_CC how_RGQ sincerely_RR he_PPHS1 was_VBDZ behaving_VVG ,_, in_II the_AT production_NN1 of_IO a_AT1 discourse_NN1 fragment_NN1 ,_, any_DD claims_NN2 regarding_II the_AT implicatures_NN2 identified_VVN will_VM have_VHI the_AT status_NN1 of_IO interpretations_NN2 ._. 
In_II this_DD1 respect_NN1 ,_, the_AT discourse_NN1 analyst_NN1 is_VBZ not_XX in_II the_AT apparently_RR secure_JJ position_NN1 of_IO the_AT formal_JJ linguist_NN1 who_PNQS has_58 '_GE rules_NN2 '_GE of_IO the_AT language_NN1 which_DDQ are_VBR or_CC are_VBR not_XX satisfied_JJ ,_, but_CCB rather_RR ,_, is_VBZ in_II the_AT position_NN1 of_IO the_AT hearer_NN1 who_PNQS has_VHZ interpretations_NN2 of_IO the_AT discourse_NN1 which_DDQ do_VD0 ,_, or_CC do_VD0 not_XX ,_, make_VV0 sense_NN1 ._. 
(_( For_IF a_AT1 more_RGR detailed_JJ treatment_NN1 of_IO conversational_JJ implicature_NN1 ,_, see_VV0 Levinson_NP1 ,_, forthcoming_JJ ._. )_) 
Inference_NN1 Since_CS the_AT discourse_NN1 analyst_NN1 ,_, like_II the_AT hearer_NN1 ,_, has_VHZ no_AT direct_JJ access_NN1 to_II a_AT1 speaker_NN1 's_VBZ intended_VVN meaning_VVG in_II producing_VVG an_AT1 utterance_NN1 ,_, he_PPHS1 often_RR has_VHZ to_TO rely_VVI on_II a_AT1 process_NN1 of_IO inference_NN1 to_TO arrive_VVI at_II an_AT1 interpretation_NN1 for_IF utterances_NN2 or_CC for_IF the_AT connections_NN2 between_II utterances_NN2 ._. 
Such_DA inferences_NN2 appear_VV0 to_TO be_VBI of_IO different_JJ kinds_NN2 ._. 
It_PPH1 may_VM be_VBI the_AT case_NN1 that_CST we_PPIS2 are_VBR capable_JJ of_IO deriving_VVG a_AT1 specific_JJ conclusion_NN1 (_( 7c_FO )_) from_II specific_JJ premises_NN2 (_( 7a_FO )_) and_CC (_( 7b_FO )_) ,_, via_II deductive_JJ inference_NN1 ,_, but_CCB we_PPIS2 are_VBR rarely_RR asked_VVN to_TO do_VDI so_RR in_II the_AT everyday_JJ discourse_NN1 we_PPIS2 encounter._NNU (_( 7_MC )_) a_AT1 ._. 
If_CS it_PPH1 's_VBZ sunny_JJ ,_, it_PPH1 's_VBZ warm._NNU b_ZZ1 ._. 
It_PPH1 's_VBZ sunny._NNU c_ZZ1 ._. 
So_RR ,_, it_PPH1 's_VBZ warm_JJ ._. 
We_PPIS2 are_VBR more_RGR likely_JJ to_TO operate_VVI with_IW a_AT1 rather_RG loose_JJ form_NN1 of_IO inferencing_VVG which_DDQ leads_VVZ us_PPIO2 to_TO believe_VVI that_CST the_AT hats_NN2 and_CC coats_NN2 mentioned_VVN in_II (_( 8_MC )_) belong_VV0 to_II visitors_NN2 to_II the_AT house_NN1 which_DDQ has_VHZ the_AT dresser_NN1 in_II its_APPGE kitchen._NNU (_( 8_MC )_) in_II the_AT kitchen_NN1 there_EX was_VBDZ a_AT1 huge_JJ dresser_NN1 and_CC when_CS anyone_PN1 went_VVD in_II you_PPY see_VV0 +_FO the_AT hats_NN2 and_CC coats_NN2 were_VBDR all_DB dumped_VVN on_II this_DD1 dresser_NN1 It_PPH1 may_VM be_VBI ,_, of_RR21 course_RR22 ,_, that_DD1 such_DA an_AT1 inference_NN1 is_VBZ wrong_JJ ,_, but_CCB ,_, as_CSA discourse_NN1 processors_NN2 ,_, we_PPIS2 seem_VV0 to_TO prefer_VVI to_TO make_VVI inferences_NN2 which_DDQ have_VH0 some_DD likelihood_NN1 of_IO being_VBG justified_VVN and_CC ,_, if_CS some_DD subsequent_JJ information_NN1 does_VDZ not_XX fit_VVI in_RP with_IW this_DD1 inference_NN1 ,_, we_PPIS2 abandon_VV0 it_PPH1 and_CC form_VV0 another_DD1 ._. 
As_II an_AT1 illustration_NN1 of_IO this_DD1 ,_, consider_VV0 the_AT following_JJ example_NN1 (_( 9_MC )_) ,_, taken_VVN from_II Sanford_NP1 &amp;_CC Garrod_NP1 (_( 1981_MC :_: 10_MC )_) :_: (_( 9_MC )_) John_NP1 was_VBDZ on_II his_APPGE way_NN1 to_II school_NN1 ._. 
If_CS we_PPIS2 were_VBDR to_TO take_VVI a_AT1 formal_JJ view_NN1 of_IO the_AT entailments_NN2 of_IO such_DA a_AT1 declarative_JJ sentence_NN1 (_( like_II that_DD1 ,_, for_REX21 example_REX22 ,_, expressed_VVN in_II Smith_NP1 &amp;_CC Wilson_NP1 ,_, 1979_MC :_: 150f._RA )_) ,_, we_PPIS2 would_VM be_VBI obliged_VVN to_TO accept_VVI as_CSA entailments_NN2 a_AT1 set_NN1 of_IO sentences_NN2 which_DDQ would_VM include_VVI the_AT following_JJ :_: (_( 10_MC )_) a_AT1 ._. 
Someone_PN1 was_VBDZ on_II his_APPGE way_NN1 to_II school._NN1 b_ZZ1 ._. 
John_NP1 was_VBDZ on_II his_APPGE way_NN1 to_II somewhere._NNU c_ZZ1 ._. 
Someone_PN1 was_VBDZ on_II his_APPGE way_NN1 to_II somewhere_RL ._. 
This_DD1 view_NN1 of_IO what_DDQ we_PPIS2 infer_VV0 from_II reading_NN1 (_( 9_MC )_) will_VM only_RR provide_VVI us_PPIO2 with_IW a_AT1 limited_JJ insight_NN1 into_II how_RRQ readers_NN2 interpret_VV0 what_DDQ they_PPHS2 read_VVD ._. 
Most_DAT readers_NN2 report_VV0 that_CST they_PPHS2 infer_VV0 from_II (_( 9_MC )_) that_CST John_NP1 is_VBZ a_AT1 schoolboy_NN1 ,_, among_II other_JJ things_NN2 ._. 
When_CS sentence_NN1 (_( 9_MC )_) is_VBZ followed_VVN later_RRR in_II the_AT same_DA text_NN1 by_II sentence_NN1 (_( II_MC )_) ,_, readers_NN2 readily_RR abandon_VV0 their_APPGE original_JJ inference_NN1 and_CC form_VV0 another_DD1 ,_, for_REX21 example_REX22 that_CST John_NP1 is_VBZ a_AT1 schoolteacher._NNU (_( 11_MC )_) Last_MD week_NNT1 he_PPHS1 had_VHD been_VBN unable_JK to_TO control_VVI the_AT class_NN1 ._. 
In_BCL21 order_BCL22 to_TO capture_VVI this_DD1 type_NN1 of_IO inference_NN1 ,_, which_DDQ is_VBZ extremely_RR common_JJ in_II our_APPGE interpretation_NN1 of_IO discourse_NN1 ,_, we_PPIS2 need_VV0 a_AT1 relatively_RR loose_JJ notion_NN1 of_IO inference_NN1 based_VVN on_II socio-cultural_JJ knowledge_NN1 ._. 
Gumperz_NP1 (_( 1977_MC )_) presents_VVZ an_AT1 extended_JJ discussion_NN1 of_IO the_AT types_NN2 of_IO factors_NN2 involved_JJ in_II this_DD1 type_NN1 of_IO pragmatic_JJ ,_, as_II31 opposed_II32 to_II33 logical_JJ ,_, inference_NN1 ._. 
We_PPIS2 shall_VM discuss_VVI the_AT influence_NN1 of_IO inference_NN1 in_II more_DAR detail_NN1 in_II Chapter_NN1 7_MC ._. 
For_IF the_AT moment_NN1 ,_, we_PPIS2 simply_RR present_VV0 a_AT1 view_NN1 which_DDQ claims_VVZ that_CST the_AT terms_NN2 reference_VV0 ,_, presupposition_NN1 ,_, implicature_NN1 and_CC inference_NN1 must_VM be_VBI treated_VVN as_CSA pragmatic_JJ concepts_NN2 in_II the_AT analysis_NN1 of_IO discourse_NN1 ._. 
These_DD2 terms_NN2 will_VM be_VBI used_VVN to_TO indicate_VVI relationships_NN2 between_II discourse_NN1 participants_NN2 and_CC elements_NN2 in_II the_AT discourse_NN1 ._. 
Since_CS the_AT pragmatic_JJ use_NN1 of_IO these_DD2 terms_NN2 is_VBZ closely_RR tied_VVN to_II the_AT context_NN1 in_II which_DDQ a_AT1 discourse_NN1 occurs_VVZ ,_, we_PPIS2 shall_VM now_RT investigate_VVI what_DDQ aspects_NN2 of_IO context_NN1 have_VH0 to_TO be_VBI considered_VVN in_II undertaking_VVG the_AT analysis_NN1 of_IO discourse_NN1 ._. 
The_AT context_NN1 of_IO situation_NN1 Since_CS the_AT beginning_NN1 of_IO the_AT 1970s_MC2 ,_, linguists_NN2 have_VH0 become_VVN increasingly_RR aware_JJ of_IO the_AT importance_NN1 of_IO context_NN1 in_II the_AT interpretation_NN1 of_IO sentences_NN2 ._. 
The_AT implications_NN2 of_IO taking_VVG context_NN1 into_II account_NN1 are_VBR well_RR expressed_VVN by_II Sadock_NP1 (_( 1978_MC :_: 281_MC )_) :_: There_EX is_VBZ ,_, then_RT ,_, a_AT1 serious_JJ methodological_JJ problem_NN1 that_CST confronts_VVZ the_AT advocate_NN1 of_IO linguistic_JJ pragmatics_NN1 ._. 
Given_VVN some_DD aspects_NN2 of_IO what_DDQ a_AT1 sentence_NN1 conveys_VVZ in_II a_AT1 particular_JJ context_NN1 ,_, is_VBZ that_DD1 aspect_NN1 part_NN1 of_IO what_DDQ the_AT sentence_NN1 conveys_VVZ in_II virtue_NN1 of_IO its_APPGE meaning_NN1 ..._... or_CC should_VM it_PPH1 be_VBI '_GE worked_VVD out_RP '_GE on_II the_AT basis_NN1 of_IO Gricean_JJ principles_NN2 from_II the_AT rest_NN1 of_IO the_AT meaning_NN1 of_IO the_AT sentence_NN1 and_CC relevant_JJ facts_NN2 of_IO the_AT context_NN1 of_IO utterance_NN1 ?_? 
If_CS we_PPIS2 are_VBR to_TO begin_VVI to_TO consider_VVI the_AT second_MD part_NN1 of_IO this_DD1 question_NN1 seriously_RR we_PPIS2 need_VV0 to_TO be_VBI able_JK to_TO specify_VVI what_DDQ are_VBR the_AT '_GE relevant_JJ facts_NN2 of_IO the_AT context_NN1 of_IO utterance_NN1 '_GE ._. 
The_AT same_DA problem_NN1 is_VBZ raised_VVN by_II Fillmore_NP1 (_( 1977_MC :_: 119_MC )_) when_RRQ he_PPHS1 advocates_NN2 a_AT1 methodology_NN1 to_II which_DDQ a_AT1 discourse_NN1 analyst_NN1 may_VM often_RR wish_VVI to_TO appeal_VVI :_: The_AT task_NN1 is_VBZ to_TO determine_VVI what_DDQ we_PPIS2 can_VM know_VVI about_II the_AT meaning_NN1 and_CC context_NN1 of_IO an_AT1 utterance_NN1 given_VVN only_RR the_AT knowledge_NN1 that_CST the_AT utterance_NN1 has_VHZ occurred_VVN ..._... 
I_PPIS1 find_VV0 that_CST whenever_RRQV I_PPIS1 notice_VV0 some_DD sentence_NN1 in_II context_NN1 ,_, I_PPIS1 immediately_RR find_VV0 myself_PPX1 asking_VVG what_DDQ the_AT effect_NN1 would_VM have_VHI been_VBN if_CS the_AT context_NN1 had_VHD been_VBN slightly_RR different_JJ ._. 
In_BCL21 order_BCL22 to_TO make_VVI appeal_NN1 to_II this_DD1 methodology_NN1 ,_, which_DDQ is_VBZ very_RG commonly_RR used_VVN in_II linguistic_JJ and_CC philosophical_JJ discussion_NN1 ,_, we_PPIS2 need_VV0 to_TO know_VVI what_DDQ it_PPH1 would_VM mean_VVI for_IF the_AT context_NN1 to_TO be_VBI '_GE slightly_RR different_JJ '_GE ._. 
Features_NN2 of_IO context_NN1 Consider_VV0 two_MC invented_VVD scenarios_NN2 in_II which_DDQ an_AT1 identical_JJ utterance_NN1 is_VBZ produced_VVN by_II two_MC distinct_JJ speakers._NNU a_AT1 speaker_NN1 :_: a_AT1 young_JJ mother_NN1 ,_, hearer_NN1 :_: her_APPGE mother-in-law_NN1 ,_, place_VV0 :_: park_NN1 ,_, by_II a_AT1 duckpond_NN1 ,_, time_NNT1 :_: sunny_JJ afternoon_NNT1 in_II September_NPM1 1962_MC ._. 
They_PPHS2 are_VBR watching_VVG the_AT young_JJ mother_NN1 's_GE two-year-old_JJ son_NN1 chasing_VVG ducks_NN2 and_CC the_AT mother-in-law_NN1 has_VHZ just_RR remarked_VVN that_CST her_APPGE son_NN1 ,_, the_AT child_NN1 's_GE father_NN1 ,_, was_VBDZ rather_RG backward_JJ at_II this_DD1 age_NN1 ._. 
The_AT young_JJ mother_NN1 says_VVZ :_: I_PPIS1 do_VD0 think_VVI Adam_NP1 's_GE quick_JJ b_ZZ1 speaker_NN1 :_: a_AT1 student_NN1 ,_, hearers_NN2 :_: a_AT1 set_NN1 of_IO students_NN2 ,_, place_VV0 :_: sitting_VVG round_RP a_AT1 coffee_NN1 table_NN1 in_II the_AT refectory_NN1 ,_, time_NNT1 :_: evening_NNT1 in_II March_NPM1 1980_MC ._. 
John_NP1 ,_, one_MC1 of_IO the_AT group_NN1 ,_, has_VHZ just_RR told_VVN a_AT1 joke_NN1 ._. 
Everyone_PN1 laughs_VVZ except_II Adam_NP1 ._. 
Then_RT Adam_NP1 laughs_VVZ ._. 
One_MC1 of_IO the_AT students_NN2 says_VVZ :_: I_PPIS1 do_VD0 think_VVI Adam_NP1 's_GE quick_JJ (_( In_II each_DD1 case_NN1 phonological_JJ prominence_NN1 is_VBZ placed_VVN on_II Adam_NP1 ._. )_) 
Clearly_RR we_PPIS2 can_VM do_VDI a_AT1 formal_JJ analysis_NN1 on_II these_DD2 tokens_NN2 and_CC ,_, in_II both_DB2 cases_NN2 ,_, the_AT speaker_NN1 says_VVZ of_IO Adam_NP1 that_CST he_PPHS1 is_VBZ quick_JJ ._. 
It_PPH1 is_VBZ clear_JJ ,_, however_RR ,_, that_CST the_AT utterances_NN2 in_II the_AT contexts_NN2 of_IO situation_NN1 in_II which_DDQ they_PPHS2 are_VBR cited_VVN ,_, would_VM be_VBI taken_VVN to_TO convey_VVI very_RG different_JJ messages_NN2 ._. 
In_II (_( a_ZZ1 )_) we_PPIS2 shall_VM simplistically_RR assume_VVI that_CST the_AT referents_NN2 of_IO I_ZZ1 and_CC Adam_NP1 are_VBR fixed_VVN by_II spatio-temporal_JJ co-ordinates_NN2 ._. 
This_58 '_GE Adam_NP1 '_GE is_VBZ being_VBG compared_VVN (_( or_CC contrasted_VVD )_) ,_, favourably_RR ,_, with_IW his_APPGE father_NN1 ._. 
Quick_JJ ,_, may_VM be_VBI interpreted_VVN ,_, in_II the_AT context_NN1 of_IO backward_JJ ,_, as_CSA meaning_VVG something_PN1 like_II '_GE quick_JJ in_II developing_NN1 '_GE ._. 
In_II (_( b_ZZ1 )_) different_JJ referents_NN2 for_CS I_PPIS1 and_CC Adam_NP1 are_VBR fixed_VVN spatiotemporally_RR ._. 
This_58 '_GE Adam_NP1 '_GE is_VBZ being_VBG compared_VVN (_( or_CC contrasted_VVD )_) not_XX with_IW his_APPGE father_NN1 and_CC favourably_RR ,_, but_CCB with_IW the_AT set_NN1 of_IO other_JJ students_NN2 unfavourably_RR ._. 
In_II this_DD1 case_NN1 quick_RR must_VM be_VBI interpreted_VVN as_CSA meaning_VVG something_PN1 like_II '_GE quick_JJ to_TO understand_VVI /_FO react_VV0 /_FO see_VV0 the_AT joke_NN1 '_GE ._. 
Moreover_RR ,_, since_CS it_PPH1 is_VBZ said_VVN in_II a_AT1 context_NN1 where_RRQ Adam_NP1 has_VHZ just_RR manifestly_RR failed_VVN to_TO react_VVI to_II the_AT punch-line_NN1 as_RG quickly_RR as_CSA the_AT set_NN1 of_IO other_JJ students_NN2 ,_, the_AT speaker_NN1 (_( given_VVN this_DD1 type_NN1 of_IO speaker_NN1 to_II this_DD1 type_NN1 of_IO hearer_NN1 in_II this_DD1 type_NN1 of_IO surroundings_NN2 )_) will_VM be_VBI assumed_VVN not_XX to_TO be_VBI intending_VVG to_TO tell_VVI an_AT1 untruth_NN1 ,_, but_CCB to_TO be_VBI implicating_VVG the_AT opposite_NN1 of_IO what_DDQ she_PPHS1 has_VHZ said_VVN ._. 
Is_VBZ it_PPH1 possible_JJ to_TO determine_VVI in_II any_DD principled_JJ way_NN1 what_DDQ aspects_NN2 of_IO context_NN1 of_IO situation_NN1 are_VBR relevant_JJ to_II these_DD2 different_JJ interpretations_NN2 of_IO the_AT same_DA '_GE utterance_NN1 '_GE on_II two_MC occasions_NN2 ?_? 
J._NP1 R._NP1 Firth_NN1 (_( regarded_VVN by_II many_DA2 as_II the_AT founder_NN1 of_IO modern_JJ British_JJ linguistics_NN1 )_) remarked_VVD :_: Logicians_NN2 are_VBR apt_JJ to_TO think_VVI of_IO words_NN2 and_CC propositions_NN2 as_CSA having_VHG '_GE meaning_NN1 '_GE somehow_RR in_II themselves_PPX2 ,_, apart_II21 from_II22 participants_NN2 in_II contexts_NN2 of_IO situation_NN1 ._. 
Speakers_NN2 and_CC listeners_NN2 do_VD0 not_XX seem_VVI to_TO be_VBI necessary_JJ ._. 
I_PPIS1 suggest_VV0 that_CST voices_NN2 should_VM not_XX be_VBI entirely_RR dissociated_VVN from_II the_AT social_JJ context_NN1 in_II which_DDQ they_PPHS2 function_VV0 and_CC that_CST therefore_RR all_DB texts_NN2 in_II modern_JJ spoken_JJ languages_NN2 should_VM be_VBI regarded_VVN as_II having_VHG '_GE the_AT implication_NN1 of_IO utterance_NN1 '_GE ,_, and_CC be_VBI referred_VVN to_II typical_JJ participants_NN2 in_II some_DD generalised_JJ context_NN1 of_IO situation._NNU (_( 1957_MC :_: 226_MC )_) Firth_NN1 ,_, then_RT ,_, was_VBDZ concerned_JJ to_TO embed_VVI the_AT utterance_NN1 in_II the_AT '_GE social_JJ context_NN1 '_GE and_CC to_TO generalise_VVI across_II meanings_NN2 in_II specified_JJ social_JJ contexts_NN2 ._. 
He_PPHS1 proposed_VVD an_AT1 approach_NN1 to_II the_AT principled_JJ description_NN1 of_IO such_DA contexts_NN2 which_DDQ bears_VVZ a_AT1 close_JJ resemblance_NN1 to_II more_RGR recent_JJ descriptions_NN2 which_DDQ we_PPIS2 shall_VM go_VVI on_RP to_TO examine_VVI :_: My_APPGE view_NN1 was_VBDZ ,_, and_CC still_RR is_VBZ ,_, that_DD1 '_VBZ context_NN1 of_IO situation_NN1 '_GE is_VBZ best_RRT used_VVN as_II a_AT1 suitable_JJ schematic_JJ construct_VV0 to_TO apply_VVI to_II language_NN1 events_NN2 ..._... 
A_AT1 context_NN1 of_IO situation_NN1 for_IF linguistic_JJ work_NN1 brings_VVZ into_II relation_NN1 the_AT following_JJ categories_NN2 :_: A_ZZ1 The_AT relevant_JJ features_NN2 of_IO participants_NN2 :_: persons_NN2 ,_, personalities._NNU i_ZZ1 The_AT verbal_JJ action_NN1 of_IO the_AT participants._NNU ii_MC The_AT non-verbal_JJ action_NN1 of_IO the_AT participants_NN2 ._. 
B_ZZ1 The_AT relevant_JJ objects_NN2 ._. 
C_NP1 The_AT effect_NN1 of_IO the_AT verbal_JJ action_NN1 ._. 
A_AT1 very_RG rough_JJ parallel_NN1 to_II this_DD1 sort_NN1 of_IO context_NN1 can_VM be_VBI found_VVN in_II language_NN1 manuals_NN2 providing_VVG the_AT learner_NN1 with_IW a_AT1 picture_NN1 of_IO the_AT railway_NN1 station_NN1 and_CC the_AT operative_JJ words_NN2 for_IF travelling_VVG by_II train_NN1 ._. 
It_PPH1 is_VBZ very_RG rough_JJ ._. 
But_CCB it_PPH1 is_VBZ parallel_RR with_IW the_AT grammatical_JJ rules_NN2 ,_, and_CC is_VBZ based_VVN on_II the_AT repetitive_JJ routines_NN2 of_IO initiated_JJ persons_NN2 in_II the_AT society_NN1 under_RG description._NNU (_( 1957_MC :_: 182_MC ;_; for_IF a_AT1 practical_JJ application_NN1 of_IO Firth_NN1 's_GE approach_NN1 ,_, see_VV0 Mitchell_NP1 ,_, 1957_MC ._. )_) 
An_AT1 approach_NN1 similarly_RR emphasising_VVG the_AT importance_NN1 of_IO an_AT1 ethnographic_JJ view_NN1 of_IO communicative_JJ events_NN2 within_II communities_NN2 has_VHZ been_VBN developed_VVN by_II Hymes_NP1 in_II a_AT1 series_NN of_IO articles_NN2 ._. 
Hymes_NP1 views_VVZ the_AT role_NN1 of_IO context_NN1 in_II interpretation_NN1 as_CSA ,_, on_II the_AT one_MC1 hand_NN1 ,_, limiting_VVG the_AT range_NN1 of_IO possible_JJ interpretations_NN2 and_CC ,_, on_II the_AT other_JJ ,_, as_CSA supporting_VVG the_AT intended_JJ interpretation_NN1 :_: The_AT use_NN1 of_IO a_AT1 linguistic_JJ form_NN1 identifies_VVZ a_AT1 range_NN1 of_IO meanings_NN2 ._. 
A_AT1 context_NN1 can_VM support_VVI a_AT1 range_NN1 of_IO meanings_NN2 ._. 
When_CS a_AT1 form_NN1 is_VBZ used_VVN in_II a_AT1 context_NN1 it_PPH1 eliminates_VVZ the_AT meanings_NN2 possible_JJ to_II that_DD1 context_NN1 other_II21 than_II22 those_DD2 the_AT form_NN1 can_VM signal_VVI :_: the_AT context_NN1 eliminates_VVZ from_II consideration_NN1 the_AT meanings_NN2 possible_JJ to_II the_AT form_NN1 other_II21 than_II22 those_DD2 the_AT context_NN1 can_VM support_VVI ._. 
(_( Hymes_NP1 ,_, 1962_MC ,_, quoted_VVN in_II Wootton_NP1 ,_, 1975_MC :_: 44_MC )_) Hymes_NP1 (_( 1964_MC )_) sets_VVZ about_II specifying_VVG the_AT features_NN2 of_IO context_NN1 which_DDQ may_VM be_VBI relevant_JJ to_II the_AT identification_NN1 of_IO a_AT1 type_NN1 of_IO speech_NN1 event_NN1 in_II a_AT1 way_NN1 reminiscent_JJ of_IO Firth_NN1 's_GE ._. 
Like_II Firth_NN1 ,_, he_PPHS1 seizes_VVZ first_MD on_II the_AT '_GE persons_NN2 '_GE participating_JJ in_II the_AT speech_NN1 event_NN1 ._. 
Generalising_VVG over_II speech_NN1 events_NN2 ,_, he_PPHS1 abstracts_NN2 the_AT roles_NN2 addressor_NN1 and_CC addressee_NN1 ._. 
The_AT addressor_NN1 is_VBZ the_AT speaker_NN1 or_CC writer_NN1 who_PNQS produces_VVZ the_AT utterance_NN1 ._. 
The_AT addressee_NN1 is_VBZ the_AT hearer_NN1 or_CC reader_NN1 who_PNQS is_VBZ the_AT recipient_NN1 of_IO the_AT utterance_NN1 ._. 
(_( Later_JJR Hymes_NN2 also_RR distinguishes_VVZ audience_NN1 ,_, since_CS the_AT presence_NN1 of_IO overhearers_NN2 may_VM contribute_VVI to_II the_AT specification_NN1 of_IO the_AT speech_NN1 event_NN1 ._. )_) 
Knowledge_NN1 of_IO the_AT addressor_NN1 in_II a_AT1 given_JJ communicative_JJ event_NN1 makes_VVZ it_PPH1 possible_JJ for_IF the_AT analyst_NN1 to_TO imagine_VVI what_DDQ that_DD1 particular_JJ person_NN1 is_VBZ likely_JJ to_TO say_VVI ._. 
Knowledge_NN1 of_IO his_APPGE addressee_NN1 constrains_VVZ the_AT analyst_NN1 's_GE expectations_NN2 even_RR further_RRR ._. 
Thus_RR ,_, if_CS you_PPY know_VV0 the_AT speaker_NN1 is_VBZ the_AT prime_JJ minister_NN1 or_CC the_AT departmental_JJ secretary_NN1 or_CC your_APPGE family_NN1 doctor_NN1 or_CC your_APPGE mother_NN1 ,_, and_CC you_PPY know_VV0 that_CST the_AT speaker_NN1 is_VBZ speaking_VVG to_II a_AT1 colleague_NN1 or_CC his_APPGE bank_NN1 manager_NN1 or_CC a_AT1 small_JJ child_NN1 ,_, you_PPY will_VM have_VHI different_JJ expectations_NN2 of_IO the_AT sort_NN1 of_IO language_NN1 which_DDQ will_VM be_VBI produced_VVN ,_, both_RR with_II31 respect_II32 to_II33 form_NN1 and_CC to_TO content_VVI ._. 
If_CS you_PPY know_VV0 ,_, further_RRR ,_, what_DDQ is_VBZ being_VBG talked_VVN about_II ,_, Hymes_NP1 '_GE category_NN1 of_IO topic_NN1 ,_, your_APPGE expectations_NN2 will_VM be_VBI further_RRR constrained_VVN ._. 
If_CS then_RT you_PPY have_VH0 information_NN1 about_II the_AT setting_NN1 ,_, both_RR in_II31 terms_II32 of_II33 where_RRQ the_AT event_NN1 is_VBZ situated_VVN in_II place_NN1 and_CC time_NNT1 ,_, and_CC in_II31 terms_II32 of_II33 the_AT physical_JJ relations_NN2 of_IO the_AT interactants_NN2 with_II31 respect_II32 to_II33 posture_NN1 and_CC gesture_NN1 and_CC facial_JJ expression_NN1 ,_, your_APPGE expectations_NN2 will_VM be_VBI still_RR further_RRR limited_VVN ._. 
The_AT remaining_JJ features_NN2 of_IO context_NN1 which_DDQ Hymes_NP1 discusses_VVZ (_( in_II 1964_MC )_) include_VV0 large-scale_JJ features_NN2 like_II channel_NN1 (_( how_RRQ is_VBZ contact_NN1 between_II the_AT participants_NN2 in_II the_AT event_NN1 being_VBG maintained_VVN by_II speech_NN1 ,_, writing_VVG ,_, signing_VVG ,_, smoke_VV0 signals_NN2 )_) ,_, code_NN1 (_( what_DDQ language_NN1 ,_, or_CC dialect_NN1 ,_, or_CC style_NN1 of_IO language_NN1 is_VBZ being_VBG used_VVN )_) ,_, message-form_NN1 (_( what_DDQ form_VV0 is_VBZ intended_VVN chat_NN1 ,_, debate_NN1 ,_, sermon_NN1 ,_, fairy-tale_NN1 ,_, sonnet_NN1 ,_, love-letter_NN1 ,_, etc._RA )_) and_CC event_NN1 (_( the_AT nature_NN1 of_IO the_AT communicative_JJ event_NN1 within_II which_DDQ a_AT1 genre_NN1 may_VM be_VBI embedded_VVN thus_RR a_AT1 sermon_NN1 or_CC prayer_NN1 may_VM be_VBI part_NN1 of_IO the_AT larger_JJR event_NN1 ,_, a_AT1 church_NN1 service_NN1 )_) ._. 
In_II later_JJR recensions_NN2 Hymes_NP1 adds_VVZ other_JJ features_NN2 ,_, for_REX21 example_REX22 key_NN1 (_( which_DDQ involves_VVZ evaluation_NN1 was_VBDZ it_PPH1 a_AT1 good_JJ sermon_NN1 ,_, a_AT1 pathetic_JJ explanation_NN1 ,_, etc._RA )_) ,_, and_CC purpose_NN1 (_( what_DDQ did_VDD the_AT participants_NN2 intend_VVI should_VM come_VVI about_RP as_II a_AT1 result_NN1 of_IO the_AT communicative_JJ event_NN1 )_) ._. 
Hymes_NP1 intends_VVZ that_CST these_DD2 contextual_JJ features_NN2 should_VM be_VBI regarded_VVN rather_RR as_CSA general_JJ phonetic_JJ features_NN2 are_VBR regarded_VVN ._. 
Just_RR as_II a_AT1 phonetician_NN1 may_VM select_VVI ,_, from_II the_AT general_JJ phonetic_JJ features_NN2 available_JJ ,_, the_AT features_NN2 voiced_VVD ,_, bilabial_JJ and_CC stop_VV0 ,_, but_CCB not_XX lateral_JJ ,_, to_TO characterise_VVI a_AT1 &lsqb;_( b_ZZ1 &rsqb;_) ,_, so_RR ,_, he_PPHS1 suggests_VVZ ,_, the_AT analyst_NN1 may_VM choose_VVI from_II the_AT contextual_JJ features_NN2 ,_, those_DD2 necessary_JJ to_TO characterise_VVI a_AT1 particular_JJ communicative_JJ event_NN1 ._. 
Just_RR as_CSA the_AT phonetician_NN1 may_VM wish_VVI to_TO make_VVI a_AT1 more_RGR detailed_JJ ,_, more_RGR specific_JJ description_NN1 of_IO the_AT &lsqb;_( b_ZZ1 &rsqb;_) under_II consideration_NN1 ,_, for_REX21 example_REX22 mentioning_VVG delayed_JJ onset_NN1 of_IO voicing_VVG and_CC some_DD protrusion_NN1 of_IO the_AT lips_NN2 during_II the_AT period_NN1 of_IO closure_NN1 ,_, so_RR may_VM the_AT ethnographer_NN1 wish_VVI to_TO specify_VVI some_DD of_IO the_AT contextual_JJ features_NN2 in_II great_JJ detail_NN1 ._. 
We_PPIS2 shall_VM return_VVI to_II this_DD1 point_NN1 ._. 
Hymes_NP1 '_GE features_NN2 constitute_VV0 essentially_RR a_AT1 checklist_NN1 which_DDQ would_VM enable_VVI a_AT1 visiting_JJ ethnographer_NN1 to_TO arrive_VVI by_II helicopter_NN1 in_II a_AT1 location_NN1 where_RRQ a_AT1 communicative_JJ event_NN1 is_VBZ in_II process_NN1 and_CC to_TO check_VVI off_II the_AT detail_NN1 of_IO the_AT nature_NN1 of_IO the_AT communicative_JJ event_NN1 ._. 
Let_VV0 us_PPIO2 consider_VVI such_DA an_AT1 ethnographer_NN1 as_CSA an_AT1 invisible_JJ witness_NN1 to_II a_AT1 particular_JJ speech_NN1 event_NN1 ._. 
He_PPHS1 would_VM begin_VVI ,_, presumably_RR ,_, by_II noting_VVG the_AT larger-scale_JJ features_NN2 of_IO context_NN1 :_: what_DDQ channel_NN1 is_VBZ being_VBG used_VVN (_( we_PPIS2 shall_VM say_VVI speech_NN1 )_) ,_, what_DDQ language_NN1 code_NN1 is_VBZ being_VBG used_VVN (_( we_PPIS2 shall_VM specify_VVI it_PPH1 is_VBZ English_JJ )_) ,_, what_DDQ message-form_NN1 is_VBZ being_VBG performed_VVN (_( we_PPIS2 shall_VM specify_VVI it_PPH1 is_VBZ conversation_NN1 )_) ,_, what_DDQ event_NN1 is_VBZ it_PPH1 embedded_VVN in_II (_( we_PPIS2 shall_VM specify_VVI it_PPH1 is_VBZ part_NN1 of_IO an_AT1 interview_NN1 )_) ._. 
He_PPHS1 can_VM identify_VVI the_AT participants_NN2 :_: the_AT addressor_NN1 is_VBZ a_AT1 young_JJ scientist_NN1 who_PNQS is_VBZ being_VBG interviewed_VVN by_II the_AT addressee_NN1 who_PNQS is_VBZ doing_VDG research_NN1 on_II language_NN1 ._. 
The_AT setting_NN1 is_VBZ physically_RR located_VVN in_II the_AT addressee_NN1 's_GE territory_NN1 in_II Edinburgh_NP1 University_NN1 and_CC a_AT1 prominent_JJ physical_JJ feature_NN1 is_VBZ a_AT1 tape-recorder_NN1 which_DDQ is_VBZ switched_VVN on_RP ._. 
The_AT time_NNT1 is_VBZ during_II the_AT later_JJR 1970s_MC2 (_( so_CS it_PPH1 is_VBZ reasonable_JJ to_TO expect_VVI that_CST they_PPHS2 will_VM speak_VVI modern_JJ English_NN1 ,_, with_IW Scottish_JJ accents_NN2 )_) ._. 
It_PPH1 has_VHZ just_RR been_VBN agreed_VVN that_CST they_PPHS2 will_VM talk_VVI about_II the_AT young_JJ scientist_NN1 's_GE work_NN1 ,_, the_AT tape-recorder_NN1 is_VBZ switched_VVN on_RP and_CC he_PPHS1 says_VVZ :_: (_( 12_MC )_) I_PPIS1 must_VM admit_VVI I_PPIS1 'm_VBM very_RG nervous_JJ ._. 
His_APPGE topic_NN1 at_II this_DD1 point_NN1 ,_, we_PPIS2 shall_VM simplistically_RR assume_VVI (_( see_VV0 further_JJR discussion_NN1 in_II Chapter_NN1 3_MC )_) ,_, is_VBZ his_APPGE nervousness_NN1 ._. 
Given_VVN the_AT knowledge_NN1 of_IO context_NN1 the_AT analyst_NN1 has_VHZ ,_, he_PPHS1 should_VM find_VVI this_DD1 a_AT1 fairly_RR unsurprising_VVG utterance_NN1 ._. 
It_PPH1 is_VBZ very_RG rarely_RR the_AT case_NN1 in_II real_JJ life_NN1 that_CST we_PPIS2 can_VM predict_VVI in_II detail_NN1 the_AT form_NN1 and_CC content_NN1 of_IO the_AT language_NN1 which_DDQ we_PPIS2 will_VM encounter_VVI ,_, but_CCB ,_, given_VVN all_DB of_IO the_AT ethnographic_JJ information_NN1 we_PPIS2 have_VH0 specified_VVN ,_, the_AT actual_JJ occurring_JJ utterance_NN1 is_VBZ much_RR more_RGR likely_JJ (_( hence_RR ,_, we_PPIS2 assume_VV0 ,_, much_RR more_RGR readily_RR processed_VVN by_II the_AT addressee_NN1 )_) than_CSN any_DD of_IO the_AT following_JJ '_GE utterances_NN2 '_GE which_DDQ did_VDD not_XX occur_VVI :_: (_( 13_MC )_) a_AT1 ._. 
Please_RR pass_VV0 the_AT marmalade._NNU b_ZZ1 ._. 
My_APPGE cat_NN1 has_VHZ just_RR been_VBN sick_JJ again._NNU c_ZZ1 ._. 
Get_VV0 into_II the_AT box._NNU d_ZZ1 ._. 
I_PPIS1 am_VBM about_RPK to_TO make_VVI the_AT first_MD incision_NN1 ._. 
The_AT more_RRR the_AT analyst_NN1 knows_VVZ about_II the_AT features_NN2 of_IO context_NN1 ,_, the_AT more_RGR likely_JJ he_PPHS1 is_VBZ to_TO be_VBI able_JK to_TO predict_VVI what_DDQ is_VBZ likely_JJ to_TO be_VBI said_VVN (_( see_VV0 2.4_MC )_) ._. 
It_PPH1 is_VBZ further_RRR the_AT case_NN1 that_CST the_AT ethnographic_JJ features_NN2 will_VM give_VVI us_PPIO2 a_AT1 value_NN1 for_IF the_AT deictic_JJ forms_NN2 occurring_VVG in_II the_AT utterance_NN1 which_DDQ was_VBDZ actually_RR produced_VVN ._. 
Thus_RR 1_MC1 ,_, must_VM ,_, and_CC am_RA must_VM be_VBI interpreted_VVN with_II31 respect_II32 to_II33 the_AT speaker_NN1 ,_, the_AT young_JJ scientist_NN1 ,_, at_II the_AT time_NNT1 of_IO making_VVG the_AT utterance_NN1 ._. 
(_( The_AT context_NN1 here_RL makes_VVZ the_AT other_JJ possible_JJ reading_NN1 ,_, that_CST the_AT speaker_NN1 is_VBZ characteristically_RR nervous_JJ all_DB of_IO the_AT time_NNT1 ,_, so_RG unlikely_JJ as_CSA not_XX to_TO be_VBI considered_VVN apparently_RR by_II the_AT addressee_NN1 ,_, or_CC indeed_RR by_II the_AT analyst_NN1 until_CS the_AT process_NN1 of_IO analysis_NN1 was_VBDZ brought_VVN to_II conscious_JJ attention_NN1 ._. )_) 
In_II 2_MC ._. 
I_ZZ1 we_PPIS2 pointed_VVD out_RP that_DD1 deictic_JJ elements_NN2 of_IO the_AT utterances_NN2 can_VM only_RR be_VBI interpreted_VVN with_II31 respect_II32 to_II33 the_AT context_NN1 in_II which_DDQ they_PPHS2 are_VBR uttered_VVN ._. 
Hymes_NP1 '_GE checklist_NN1 of_IO ethnographic_JJ features_NN2 offers_VVZ one_MC1 characterisation_NN1 of_IO context_NN1 to_II which_DDQ we_PPIS2 can_VM relate_VVI such_DA deictic_JJ elements_NN2 ._. 
A_AT1 more_RGR elaborate_JJ checklist_NN1 is_VBZ provided_VVN by_II the_AT philosopher_NN1 Lewis_NP1 (_( 1972_MC )_) ,_, specifically_RR to_TO provide_VVI an_AT1 index_NN1 of_IO those_DD2 co-ordinates_NN2 which_DDQ a_AT1 hearer_NN1 would_VM need_VVI to_TO have_VHI specified_VVN in_BCL21 order_BCL22 that_CST he_PPHS1 could_VM determine_VVI the_AT truth_NN1 of_IO a_AT1 sentence_NN1 ._. 
Like_II most_DAT formal_JJ linguists_NN2 ,_, Lewis_NP1 assumes_VVZ that_CST the_AT channel_NN1 is_VBZ speech_NN1 ,_, the_AT code_NN1 ,_, English_NN1 ,_, the_AT message-form_JJ conversation_NN1 and_CC the_AT event_NN1 one_PN1 where_CS one_MC1 individual_NN1 is_VBZ informing_VVG another_DD1 ._. 
His_APPGE interests_NN2 lie_VV0 ,_, not_XX with_IW these_DD2 general_JJ features_NN2 of_IO the_AT communicative_JJ event_NN1 ,_, but_CCB with_IW those_DD2 particular_JJ co-ordinates_NN2 which_DDQ constitute_VV0 '_GE a_AT1 package_NN1 of_IO relevant_JJ factors_NN2 ,_, an_AT1 index_NN1 '_GE (_( 1972_MC :_: 173_MC )_) and_CC which_DDQ characterise_VV0 the_AT context_NN1 against_II which_DDQ the_AT truth_NN1 of_IO a_AT1 sentence_NN1 is_VBZ to_TO be_VBI judged_VVN ._. 
The_AT co-ordinates_NN2 of_IO the_AT index_NN1 are_VBR specified_VVN as_CSA follows_VVZ :_: a_AT1 possible-world_JJ co-ordinate_NN1 :_: this_DD1 is_VBZ to_TO account_VVI for_IF states_NN2 of_IO affairs_NN2 which_DDQ might_VM be_VBI ,_, or_CC could_VM be_VBI supposed_JJ to_TO be_VBI or_CC are_VBR b_ZZ1 time_NNT1 co-ordinate_VV0 :_: to_TO account_VVI for_IF tensed_JJ sentences_NN2 and_CC adverbials_NN2 like_II today_RT or_CC next_MD week_NNT1 c_ZZ1 place_VV0 co-ordinate_VV0 :_: to_TO account_VVI for_IF sentences_NN2 like_II here_RL it_PPH1 is_VBZ d_ZZ1 speaker_NN1 co-ordinate_VV0 :_: to_TO account_VVI for_IF sentences_NN2 which_DDQ include_VV0 first_MD person_NN1 reference_NN1 (_( I_ZZ1 ,_, me_PPIO1 ,_, we_PPIS2 ,_, our_APPGE ,_, etc._RA )_) e_ZZ1 audience_NN1 co-ordinate_VV0 :_: to_TO account_VVI for_IF sentences_NN2 including_II you_PPY ,_, yours_PPGE ,_, yourself_PPX1 ,_, etc._RA f_RA indicated_VVD object_NN1 co-ordinate_VV0 :_: to_TO account_VVI for_IF sentences_NN2 containing_VVG demonstrative_JJ phrases_NN2 like_II this_DD1 ,_, those_DD2 ,_, etc._RA g_ZZ1 previous_JJ discourse_NN1 co-ordinate_VV0 :_: to_TO account_VVI for_IF sentences_NN2 including_II phrases_NN2 like_II the_AT latter_DA ,_, the_AT aforementioned_JJ ,_, etc._RA h_ZZ1 assignment_NN1 co-ordinate_VV0 :_: an_AT1 infinite_JJ series_NN of_IO things_NN2 (_( sets_NN2 of_IO things_NN2 ,_, sequences_NN2 of_IO things_NN2 ..._... )_) 
Rather_RG similar_JJ lists_NN2 are_VBR proposed_VVN by_II scholars_NN2 who_PNQS are_VBR concerned_JJ with_IW the_AT construction_NN1 of_IO formal_JJ discourse_NN1 domains_NN2 (_( see_VV0 discussion_NN1 in_II Chapter_NN1 3_MC )_) ._. 
For_IF our_APPGE present_JJ purposes_NN2 we_PPIS2 should_VM note_VVI that_CST Lewis_NP1 '_GE list_NN1 ,_, like_II Hymes_NP1 '_GE ,_, makes_VVZ reference_NN1 to_II the_AT speaker_NN1 and_CC hearer_NN1 in_BCL21 order_BCL22 to_TO assign_VVI values_NN2 to_II the_AT deictic_JJ categories_NN2 of_IO speaker_NN1 and_CC audience_NN1 (_( addressor_NN1 /_FO addressee_NN1 )_) realised_VVD in_II first_MD and_CC second_MD person_NN1 pronouns_NN2 ._. 
Hymes_NP1 '_GE category_NN1 setting_NN1 is_VBZ expanded_VVN to_TO take_VVI explicit_JJ and_CC distinct_JJ account_NN1 of_IO time_NNT1 and_CC place_NN1 ._. 
Hymes_NP1 '_GE generalised_JJ feature_NN1 of_IO topic_NN1 is_VBZ now_RT distributed_VVN between_II the_AT deictic_JJ co-ordinate_NN1 indicated_VVD object_NN1 ,_, the_AT assignment_NN1 co-ordinate_NN1 and_CC the_AT previous_JJ discourse_NN1 co-ordinate_NN1 ._. 
This_DD1 last_MD co-ordinate_VV0 specifically_RR enables_VVZ the_AT hearer_NN1 to_TO interpret_VVI what_DDQ is_VBZ said_VVN in_II41 the_II42 light_II43 of_II44 what_DDQ has_VHZ already_RR been_VBN said_VVN ._. 
It_PPH1 builds_VVZ in_II a_AT1 cumulative_JJ temporal_JJ structure_NN1 to_II the_AT index_NN1 ,_, in_CS21 that_CS22 the_AT hearer_NN1 must_VM continually_RR update_VVI the_AT information_NN1 in_II his_APPGE previous_JJ discourse_NN1 component_NN1 ,_, to_TO take_VVI account_NN1 of_IO what_DDQ has_VHZ most_RGT recently_RR been_VBN added_VVN ._. 
It_PPH1 is_VBZ ,_, obviously_RR ,_, not_XX possible_JJ for_IF us_PPIO2 in_II a_AT1 textbook_NN1 to_TO permit_VVI you_PPY to_TO have_VHI the_AT experience_NN1 of_IO everyday_JJ discourse_NN1 in_II what_DDQ Stenning_NP1 (_( 1978_MC )_) calls_VVZ a_AT1 '_GE normal_JJ context_NN1 '_GE ,_, where_CS the_AT hearer_NN1 is_VBZ part_NN1 of_IO the_AT context_NN1 and_CC then_RT experiences_VVZ the_AT text_NN1 ._. 
We_PPIS2 have_VH0 to_TO have_VHI recourse_NN1 to_II what_DDQ Stenning_NP1 calls_NN2 '_GE abnormal_JJ '_GE contexts_NN2 ,_, where_CS the_AT analyst_NN1 reads_VVZ the_AT text_NN1 and_CC then_RT has_VHZ to_TO try_VVI to_TO provide_VVI the_AT characteristics_NN2 of_IO the_AT context_NN1 in_II which_DDQ the_AT text_NN1 might_VM have_VHI occurred_VVN ._. 
We_PPIS2 are_VBR going_VVGK to_TO provide_VVI you_PPY with_IW three_MC written_JJ fragments_NN2 ,_, abstracted_VVN from_II the_AT contexts_NN2 in_II which_DDQ they_PPHS2 appeared_VVD ._. 
The_AT first_MD two_MC are_VBR printed_VVN ,_, the_AT third_MD spraygunned_VVD on_II a_AT1 wall_NN1 ._. 
We_PPIS2 ask_VV0 you_PPY to_TO consider_VVI what_DDQ ,_, if_CS any_DD ,_, difficulty_NN1 you_PPY have_VH0 in_II understanding_VVG them_PPHO2 ,_, in_II31 terms_II32 of_II33 the_AT co-ordinates_NN2 of_IO Lewis_NP1 '_GE index._NNU (_( 14_MC )_) ._. 
Place_VV0 two_MC fingers_NN2 in_II the_AT two_MC holes_NN2 directly_RR to_II the_AT left_NN1 of_IO the_AT finger_NN1 stop_NN1 ._. 
Remove_VV0 finger_NN1 nearest_II stop._NNU b_ZZ1 ._. 
He_PPHS1 seemed_VVD to_TO resent_VVI them_PPHO2 on_II that_DD1 occasion_NN1 and_CC will_VM not_XX wear_VVI them_PPHO2 today._NNU c_ZZ1 ._. 
SQUASHED_JJ INSECTS_NN2 DO_VD0 NT_NP1 BITE_VVI MAD_JJ MENTAL_JJ RULE_NN1 We_PPIS2 have_VH0 not_XX ,_, as_RR21 yet_RR22 ,_, introduced_VVD any_DD satisfactory_JJ way_NN1 of_IO handling_VVG your_APPGE experience_NN1 of_IO previous_JJ similar_JJ texts_NN2 (_( see_VV0 discussion_NN1 in_II 2.4_MC )_) ._. 
For_IF the_AT moment_NN1 we_PPIS2 shall_VM suppose_VVI that_CST you_PPY probably_RR recognise_VV0 the_AT type_NN1 of_IO writer_NN1 in_II (_( a_ZZ1 )_) as_CSA some_DD impersonal_JJ /_FO institutionalised_JJ writer_NN1 addressing_VVG a_AT1 general_JJ reader_NN1 rather_II21 than_II22 a_AT1 particular_JJ individual_NN1 (_( paying_VVG attention_NN1 to_II Place_NN1 and_CC Remove_VV0 and_CC the_AT ellipsis_NN1 in_II the_AT second_MD sentence_NN1 (_( the_AT )_) finger_NN1 nearest_II (_( to_II the_AT )_) stop_VV0 )_) ._. 
If_CS you_PPY have_VH0 difficulty_NN1 in_II interpreting_VVG this_DD1 fragment_NN1 it_PPH1 is_VBZ probably_RR partly_RR because_CS you_PPY are_VBR not_XX sure_JJ of_IO the_AT referents_NN2 of_IO the_AT expression_NN1 the_AT two_MC holes_NN2 and_CC the_AT finger_NN1 stop_NN1 ._. 
You_PPY may_VM work_VVI out_RP that_CST the_AT two_MC holes_NN2 have_VH0 to_TO be_VBI of_IO a_AT1 suitable_JJ size_NN1 for_IF an_AT1 individual_NN1 (_( ?_? )_) to_TO put_VVI two_MC fingers_NN2 in_RP ,_, possibly_RR near_RR enough_RR to_II each_PPX221 other_PPX222 to_TO put_VVI two_MC fingers_NN2 of_IO the_AT same_DA hand_NN1 in_RP ,_, and_CC ,_, having_VHG established_VVN this_DD1 scale_NN1 ,_, it_PPH1 seems_VVZ likely_JJ that_CST the_AT object_NN1 referred_VVN to_II as_II the_AT finger_NN1 stop_NN1 is_VBZ only_JJ centimetres_NNU2 removed_VVN ,_, rather_II21 than_II22 kilometres_NNU2 removed_VVN ._. 
It_PPH1 would_VM certainly_RR help_VVI you_PPY to_TO have_VHI the_AT following_JJ information_NN1 :_: The_AT addressor_NN1 is_VBZ the_AT Post_NN1 Office_NN1 ._. 
The_AT addressee_NN1 is_VBZ you_PPY as_II a_AT1 telephone_NN1 user_NN1 ._. 
You_PPY can_VM probably_RR work_VVI out_RP the_AT rest_NN1 if_CS you_PPY did_VDD not_XX know_VVI it_PPH1 already_RR ._. 
However_RRQV we_PPIS2 shall_VM spell_VVI out_RP some_DD more_DAR :_: The_AT time_NNT1 of_IO utterance_NN1 in_II clock_NN1 or_CC calendar_NN1 time_VV0 does_VDZ not_XX seem_VVI relevant_JJ ,_, but_CCB what_DDQ certainly_RR is_VBZ relevant_JJ is_VBZ that_CST you_PPY should_VM know_VVI whether_CSW this_DD1 instruction_NN1 still_RR applies_VVZ ._. 
(_( It_PPH1 does_VDZ ._. )_) 
The_AT place_NN1 of_IO the_AT original_JJ utterance_NN1 is_VBZ hardly_RR relevant_JJ but_CCB where_CS you_PPY would_VM encounter_VVI the_AT text_NN1 is_VBZ ._. 
(_( Look_VV0 in_II your_APPGE telephone_NN1 directory_NN1 ._. )_) 
The_AT possible_JJ world_NN1 that_CST is_VBZ relevant_JJ is_VBZ specified_VVN in_II the_AT previous_JJ discourse_NN1 :_: '_" It_PPH1 is_VBZ worth_II remembering_VVG how_RRQ to_TO dial_VVI 999_MC in_II darkness_NN1 or_CC smoke_NN1 ._. 
'_GE (_( We_PPIS2 should_VM point_VVI out_RP that_CST you_PPY are_VBR not_XX here_RL being_VBG asked_VVN to_TO use_VVI the_AT co-ordinates_NN2 for_IF the_AT purpose_NN1 Lewis_NP1 intended_VVD them_PPHO2 for_IF ,_, to_TO determine_VVI the_AT truth_NN1 of_IO a_AT1 sentence_NN1 ._. 
It_PPH1 is_VBZ a_AT1 matter_NN1 of_IO debate_NN1 whether_CSW truth_NN1 can_VM be_VBI assigned_VVN to_II sentences_NN2 in_II the_AT imperative_JJ form_NN1 ._. )_) 
In_II text_NN1 b_ZZ1 the_AT problem_NN1 of_IO interpretation_NN1 arises_VVZ because_II21 of_II22 not_XX knowing_VVG the_AT referents_NN2 for_IF the_AT expressions_NN2 He_PPHS1 ,_, them_PPHO2 ,_, on_II that_DD1 occasion_NN1 and_CC them_PPHO2 and_CC not_XX having_VHG a_AT1 value_NN1 to_TO fix_VVI the_AT time_NNT1 expression_NN1 today_RT ._. 
You_PPY may_VM be_VBI able_JK to_TO work_VVI out_RP that_CST He_PPHS1 refers_VVZ to_II an_AT1 animate_JJ masculine_JJ entity_NN1 ,_, the_AT subject_NN1 of_IO both_DB2 clauses_NN2 ._. 
You_PPY may_VM wonder_VVI why_RRQ it_PPH1 is_VBZ reported_VVN that_CST He_PPHS1 seemed_VVD to_TO resent_VVI them_PPHO2 ,_, which_DDQ may_VM suggest_VVI that_CST he_PPHS1 was_VBDZ unable_JK to_TO express_VVI his_APPGE own_DA resentment_NN1 ,_, which_DDQ may_VM limit_VVI your_APPGE range_NN1 of_IO potential_JJ interpretations_NN2 of_IO the_AT expression_NN1 He_PPHS1 ._. 
You_PPY note_VV0 that_CST he_PPHS1 resented_VVD them_PPHO2 ,_, where_CS them_PPHO2 is_VBZ plural_JJ ,_, and_CC you_PPY may_VM consider_VVI what_DDQ plural_JJ entity_NN1 may_VM be_VBI both_RR resented_VVN and_CC worn_VVN (_( or_CC not_XX worn_VVN )_) ._. 
This_DD1 example_NN1 has_VHZ all_DB the_AT characteristics_NN2 of_IO a_AT1 sentence_NN1 occurring_VVG within_II a_AT1 larger_JJR piece_NN1 of_IO text_NN1 ,_, and_CC illustrates_VVZ quite_RG clearly_RR the_AT need_NN1 for_IF a_AT1 '_GE previous_JJ discourse_NN1 '_GE co-ordinate_NN1 ,_, as_II31 well_II32 as_II33 the_AT more_RGR obvious_58 '_GE time_NNT1 '_GE and_CC '_GE place_NN1 '_GE co-ordinates_NN2 ._. 
This_DD1 text_NN1 appeared_VVD in_II The_AT Sporting_JJ Chronicle_NN1 on_II 4_MC June_NPM1 1980_MC ._. 
In_II the_AT preceding_JJ part_NN1 of_IO the_AT text_NN1 ,_, the_AT writer_NN1 has_VHZ been_VBN describing_VVG a_AT1 particular_JJ racehorse_NN1 (_( He_PPHS1 )_) which_DDQ had_VHD been_VBN fitted_VVN with_IW blinkers_NN2 (_( them_PPHO2 )_) for_IF its_APPGE previous_JJ race_NN1 (_( on_II that_DD1 occasion_NN1 )_) ._. 
The_AT third_MD text_NN1 ,_, c_ZZ1 ,_, offers_NN2 more_RGR thorny_JJ problems_NN2 ._. 
Whereas_CS the_AT language_NN1 of_IO a_AT1 and_CC b_ZZ1 is_VBZ quite_RG straightforward_JJ and_CC all_DB you_PPY require_VV0 to_TO arrive_VVI at_II an_AT1 interpretation_NN1 are_VBR values_NN2 for_IF expressions_NN2 being_VBG used_VVN to_TO refer_VVI ,_, you_PPY may_VM feel_VVI that_CST the_AT language_NN1 here_RL is_VBZ obscure_JJ ,_, perhaps_RR not_XX even_RR meaningful_JJ ._. 
It_PPH1 is_VBZ relevant_JJ that_CST the_AT time_NNT1 at_II which_DDQ this_DD1 text_NN1 appeared_VVD was_VBDZ in_II the_AT late_JJ 1970s_MC2 ._. 
Your_APPGE experience_NN1 of_IO previous_JJ similar_JJ texts_NN2 in_II the_AT 1970s_MC2 may_VM have_VHI familiarised_VVN you_PPY with_IW the_AT form_NN1 X_ZZ1 rule_NN1 OK_RR which_DDQ may_VM permit_VVI you_PPY to_TO divide_VVI this_DD1 unpunctuated_JJ sequence_NN1 into_II two_MC parts_NN2 :_: SQUASHED_JJ INSECTS_NN2 DO_VD0 NT_NP1 BITE_VVI MAD_JJ MENTAL_JJ RULE_NN1 The_AT place_NN1 at_II which_DDQ it_PPH1 appeared_VVD is_VBZ relevant_JJ ._. 
It_PPH1 was_VBDZ spraygunned_VVN on_II a_AT1 wall_NN1 in_II Glasgow_NP1 ._. 
The_AT form_NN1 of_IO the_AT text_NN1 ,_, together_RL with_IW the_AT information_NN1 about_II place_NN1 ,_, may_VM suggest_VVI to_II you_PPY ,_, if_CS you_PPY have_VH0 previous_JJ experience_NN1 of_IO such_DA texts_NN2 ,_, that_CST this_DD1 text_NN1 derives_VVZ from_II an_AT1 interaction_NN1 between_II street_NN1 gangs_NN2 ._. 
Encyclopedic_JJ knowledge_NN1 of_IO the_AT world_NN1 might_VM inform_VVI you_PPY that_CST the_AT writer_NN1 is_VBZ a_AT1 member_NN1 of_IO '_GE Mad_JJ Mental_JJ '_GE (_( a_AT1 street_NN1 gang_NN1 )_) and_CC that_CST the_AT intended_JJ addressees_NN2 are_VBR members_NN2 of_IO '_GE The_AT Insects_NN2 '_GE (_( another_DD1 street_NN1 gang_NN1 )_) ._. 
You_PPY then_RT need_VV0 to_TO make_VVI appeal_NN1 to_II previous_JJ discourse_NN1 in_II which_DDQ the_AT Insects_NN2 had_VHD proclaimed_VVN INSECTS_NN2 BITE_VV0 ._. 
(_( You_PPY might_VM then_RT appeal_VVI to_II your_APPGE knowledge_NN1 of_IO what_DDQ Hymes_NP1 calls_NN2 '_GE message-form_NN1 '_GE which_DDQ informs_VVZ you_PPY that_DD1 street_NN1 gang_NN1 interactions_NN2 on_II walls_NN2 consists_VVZ of_IO taunts_NN2 and_CC counter-taunts_NN2 ._. 
Thus_RR you_PPY might_VM arrive_VVI at_II an_AT1 attribution_NN1 of_IO intention_NN1 in_II the_AT warning_NN1 SQUASHED_JJ INSECTS_NN2 DO_VD0 NT_NP1 BITE_VVI and_CC the_AT straight_JJ assertion_NN1 MAD_JJ MENTAL_JJ RULE_NN1 without_IW the_AT OK_JJ tag_NN1 ,_, which_DDQ might_VM be_VBI taken_VVN to_TO invite_VVI assent_NN1 on_II41 the_II42 part_II43 of_II44 the_AT addressee_NN1 ._. )_) 
Texts_NN2 a_AT1 and_CC b_ZZ1 ,_, addressed_VVN to_II the_AT general_JJ reader_NN1 ,_, are_VBR relatively_RR accessible_JJ fragments_NN2 of_IO language_NN1 which_DDQ require_VV0 only_JJ specification_NN1 of_IO the_AT intended_JJ referents_NN2 to_TO make_VVI them_PPHO2 readily_RR interpretable_JJ ._. 
Text_NN1 c_ZZ1 is_VBZ intended_VVN for_IF specific_JJ addressees_NN2 ,_, not_XX for_IF the_AT general_JJ public_NN1 ,_, and_CC it_PPH1 is_VBZ hard_JJ for_IF the_AT general_JJ public_NN1 to_TO interpret_VVI without_IW access_NN1 to_II shared_JJ presuppositions_NN2 and_CC previous_JJ experience_NN1 which_DDQ can_VM not_XX comfortably_RR be_VBI forced_VVN into_II the_AT framework_NN1 proposed_VVN by_II Lewis_NP1 ._. 
In_BCL21 order_BCL22 to_TO take_VVI account_NN1 of_IO this_DD1 ,_, we_PPIS2 are_VBR going_VVGK to_TO need_VVI some_DD way_NN1 of_IO making_VVG appeals_NN2 to_II notions_NN2 like_NN1 '_GE shared_JJ presuppositions_NN2 '_GE ,_, 'encyclopedic_JJ knowledge'_NN1 ,_, '_GE intention_NN1 /_FO purpose_NN1 in_II uttering_NN1 '_GE and_CC '_GE experience_NN1 of_IO previous_JJ similar_JJ text_NN1 '_GE which_DDQ we_PPIS2 have_VH0 simply_RR appealed_VVN to_II in_II an_AT1 ad_JJ21 hoc_JJ22 way_NN1 in_II our_APPGE discussion_NN1 so_RG far_RR ._. 
We_PPIS2 return_VV0 to_II these_DD2 questions_NN2 in_II 2.3_MC ._. 
What_DDQ we_PPIS2 have_VH0 shown_VVN in_II this_DD1 section_NN1 is_VBZ that_CST the_AT contextual_JJ features_NN2 suggested_VVN by_II Hymes_NP1 ,_, supplemented_VVN with_IW the_AT index_NN1 of_IO co-ordinates_NN2 proposed_VVN by_II Lewis_NP1 (_( put_VV0 forward_RL ,_, remember_VV0 ,_, with_IW quite_RG different_JJ purposes_NN2 in_II mind_NN1 )_) do_VD0 enable_VVI us_PPIO2 to_TO give_VVI a_AT1 partial_JJ account_NN1 of_IO what_DDQ the_AT undifferentiated_JJ term_NN1 '_GE context_NN1 '_GE may_VM mean_VVI ._. 
From_II this_DD1 it_PPH1 follows_VVZ that_CST we_PPIS2 could_VM give_VVI some_DD account_NN1 of_IO what_DDQ it_PPH1 might_VM mean_VVI to_TO '_" change_VVI the_AT context_NN1 '_GE in_II the_AT sense_NN1 in_II which_DDQ Fillmore_NP1 (_( 1977_MC :_: 119_MC )_) envisages_VVZ this_DD1 when_CS he_PPHS1 says_58 '_GE I_PPIS1 ..._... find_VV0 myself_PPX1 asking_VVG what_DDQ the_AT effect_NN1 would_VM have_VHI been_VBN if_CS the_AT context_NN1 had_VHD been_VBN slightly_RR different_JJ ._. 
'_" We_PPIS2 could_VM reply_VVI that_CST if_CS you_PPY alter_VV0 the_AT condition_NN1 specified_VVN by_II any_DD of_IO the_AT co-ordinates_NN2 ,_, you_PPY alter_VV0 the_AT context_NN1 ._. 
At_II this_DD1 point_NN1 we_PPIS2 shall_VM consider_VVI only_RR the_AT alteration_NN1 of_IO one_MC1 co-ordinate_NN1 ,_, the_AT speaker_NN1 co-ordinate_NN1 ._. 
Obviously_RR ,_, if_CS Jane_NP1 says_VVZ I_PPIS1 'm_VBM skipping_VVG and_CC Mary_NP1 says_VVZ I_PPIS1 'm_VBM skipping_VVG we_PPIS2 observe_VV0 that_CST on_II one_MC1 occasion_NN1 it_PPH1 is_VBZ Jane_NP1 who_PNQS announces_VVZ that_CST she_PPHS1 is_VBZ skipping_VVG and_CC on_II another_DD1 it_PPH1 is_VBZ Mary_NP1 ._. 
In_II each_DD1 case_NN1 the_AT sentence_NN1 is_VBZ true_JJ if_CS the_AT person_NN1 who_PNQS spoke_VVD was_VBDZ skipping_VVG at_II the_AT time_NNT1 of_IO the_AT utterance_NN1 ._. 
However_RR ,_, if_CS we_PPIS2 are_VBR further_RRR told_VVN that_DD1 speaker_NN1 Jane_NP1 is_VBZ only_RR three_MC years_NNT2 old_JJ ,_, we_PPIS2 may_VM ,_, in_II31 addition_II32 to_II33 paying_VVG attention_NN1 to_II the_AT announcement_NN1 ,_, consider_VV0 that_CST it_PPH1 is_VBZ a_AT1 remarkable_JJ feat_NN1 for_IF a_AT1 three-year-old_JJ ._. 
Whereas_CS if_CS Mary_NP1 is_VBZ eight_MC years_NNT2 old_JJ and_CC known_VVN to_TO be_VBI an_AT1 intrepid_JJ skipper_NN1 ,_, the_AT announcement_NN1 may_VM be_VBI one_MC1 of_IO a_AT1 depressingly_RR predictable_JJ series_NN ._. 
We_PPIS2 pay_VV0 different_JJ amounts_NN2 of_IO attention_NN1 to_II the_AT announcements_NN2 and_CC react_VV0 to_II them_PPHO2 differently_RR ,_, because_CS one_MC1 aspect_NN1 of_IO the_AT context_NN1 ,_, the_AT speaker_NN1 ,_, is_VBZ significantly_RR different_JJ ._. 
Consider_VV0 the_AT following_JJ fragment_NN1 of_IO conversation_NN1 :_: (_( 15_MC )_) A_ZZ1 :_: are_VBR you_PPY often_RR here_RL B_ZZ1 :_: quite_RG often_RR +_FO about_RG once_RR a_AT1 month_NNT1 +_FO actually_RR ++_FO I_ZZ1 come_VV0 up_RP to_TO see_VVI my_APPGE children_NN2 You_PPY have_VH0 to_TO suppose_VVI of_IO B_ZZ1 that_CST B_ZZ1 is_VBZ of_IO an_AT1 age_NN1 to_TO have_VHI children_NN2 ._. 
What_DDQ we_PPIS2 are_VBR interested_JJ in_II is_VBZ the_AT different_JJ sorts_NN2 of_IO inferences_NN2 which_DDQ we_PPIS2 make_VV0 as_CSA addressees_NN2 ,_, depending_II21 on_II22 variables_NN2 like_II the_AT age_NN1 and_CC sex_NN1 of_IO the_AT speaker_NN1 ,_, as_II a_AT1 result_NN1 of_IO hearing_NN1 what_DDQ B_ZZ1 says_VVZ ._. 
Suppose_VV0 B_ZZ1 is_VBZ a_AT1 man_NN1 of_IO seventy_MC ._. 
We_PPIS2 assume_VV0 that_CST B_ZZ1 's_VBZ children_NN2 will_VM be_VBI grown-up_JJ ._. 
Nothing_PN1 particular_JJ follows_VVZ from_II the_AT fact_NN1 that_CST he_PPHS1 visits_VVZ them_PPHO2 once_RR a_AT1 month_NNT1 ,_, except_CS perhaps_RR we_PPIS2 infer_VV0 that_CST he_PPHS1 has_VHZ a_AT1 close_JJ relationship_NN1 with_IW them_PPHO2 ._. 
Suppose_VV0 the_AT speaker_NN1 is_VBZ a_AT1 young_JJ man_NN1 in_II his_APPGE thirties_MC2 ._. 
We_PPIS2 assume_VV0 that_CST children_NN2 he_PPHS1 has_VHZ will_VM be_VBI young_JJ children_NN2 ,_, children_NN2 of_IO an_AT1 age_NN1 who_PNQS usually_RR live_VV0 with_IW their_APPGE parents_NN2 ._. 
We_PPIS2 may_VM then_RT wonder_VVI why_RRQ B_ZZ1 's_VBZ children_NN2 are_VBR not_XX living_VVG with_IW their_APPGE father_NN1 ,_, wonder_VV0 whether_CSW the_AT exigences_NN2 of_IO his_APPGE professional_JJ life_NN1 ,_, or_CC of_IO his_APPGE relationship_NN1 with_IW the_AT children_NN2 's_GE mother_NN1 ,_, constrains_VVZ him_PPHO1 to_TO live_VVI apart_II21 from_II22 them_PPHO2 ._. 
Suppose_VV0 the_AT speaker_NN1 is_VBZ a_AT1 young_JJ woman_NN1 in_II her_APPGE thirties_MC2 ._. 
Again_RT we_PPIS2 assume_VV0 that_CST she_PPHS1 would_VM have_VHI young_JJ children_NN2 ,_, children_NN2 who_PNQS would_VM normally_RR be_VBI expected_VVN to_TO be_VBI living_VVG with_IW her_PPHO1 ._. 
Since_CS ,_, in_II the_AT case_NN1 of_IO the_AT parents_NN2 being_VBG separated_VVN ,_, young_JJ children_NN2 usually_RR live_VV0 with_IW their_APPGE mother_NN1 in_II our_APPGE society_NN1 ,_, we_PPIS2 might_VM infer_VVI that_CST the_AT woman_NN1 's_GE children_NN2 are_VBR in_II some_DD form_NN1 of_IO institutional_JJ or_CC educational_JJ care_NN1 ._. 
(_( In_II the_AT conversation_NN1 we_PPIS2 quote_VV0 from_II ,_, the_AT speaker_NN1 was_VBDZ a_AT1 man_NN1 in_II his_APPGE early_JJ thirties_MC2 and_CC the_AT children_NN2 were_VBDR living_VVG with_IW his_APPGE estranged_JJ wife_NN1 ,_, all_DB inferences_NN2 which_DDQ had_VHD been_VBN drawn_VVN by_II A_ZZ1 before_CS B_ZZ1 went_VVD on_RP to_TO explain_VVI that_CST this_DD1 was_VBDZ the_AT case_NN1 ._. )_) 
Observe_VV0 that_CST the_AT sorts_NN2 of_IO inferences_NN2 we_PPIS2 have_VH0 been_VBN discussing_VVG are_VBR not_XX sanctioned_VVN by_II the_AT form_NN1 of_IO language_NN1 used_VVD ._. 
The_AT different_JJ inferences_NN2 arise_VV0 because_II21 of_II22 the_AT alteration_NN1 of_IO the_AT context_NN1 ,_, in_II the_AT simple_JJ manipulation_NN1 of_IO age_NN1 and_CC sex_NN1 of_IO the_AT addressor_NN1 ._. 
It_PPH1 is_VBZ the_AT interpretation_NN1 of_IO the_AT utterance_NN1 in_II context_NN1 which_DDQ permits_VVZ the_AT hearer_NN1 to_TO draw_VVI such_DA inferences_NN2 (_( see_VV0 Chapter_NN1 7_MC for_IF further_JJR discussion_NN1 of_IO inferences_NN2 )_) ._. 
Co-text_NN1 In_II our_APPGE discussion_NN1 so_RG far_RR we_PPIS2 have_VH0 concentrated_VVN particularly_RR on_II the_AT physical_JJ context_NN1 in_II which_DDQ single_JJ utterances_NN2 are_VBR embedded_VVN and_CC we_PPIS2 have_VH0 paid_VVN rather_RG little_DA1 attention_NN1 to_II the_AT previous_JJ discourse_NN1 co-ordinate_NN1 ._. 
Lewis_NP1 introduced_VVD this_DD1 co-ordinate_NN1 to_TO take_VVI account_NN1 of_IO sentences_NN2 which_DDQ include_VV0 specific_JJ reference_NN1 to_II what_DDQ has_VHZ been_VBN mentioned_VVN before_RT as_CSA in_II phrases_NN2 like_II the_AT aforementioned_JJ ._. 
It_PPH1 is_VBZ ,_, however_RR ,_, the_AT case_NN1 that_CST any_DD sentence_NN1 other_II21 than_II22 the_AT first_MD in_II a_AT1 fragment_NN1 of_IO discourse_NN1 ,_, will_VM have_VHI the_AT whole_NN1 of_IO its_APPGE interpretation_NN1 forcibly_RR constrained_VVN by_II the_AT preceding_JJ text_NN1 ,_, not_XX just_RR those_DD2 phrases_NN2 which_DDQ obviously_RR and_CC specifically_RR refer_VV0 to_II the_AT preceding_JJ text_NN1 ,_, like_II the_AT aforementioned_JJ ._. 
Just_RR as_CSA the_AT interpretation_NN1 of_IO the_AT token_NN1 in_II the_AT child_NN1 's_GE representation_NN1 of_IO '_" without_RL to_TO disturb_VVI the_AT lion_NN1 '_GE and_CC the_AT token_NN1 &lsqb;_( p_ZZ1 &rsqb;_) in_II &lsqb;_( greipbritn_NN1 &rsqb;_) are_VBR determined_VVN by_II the_AT context_NN1 in_II which_DDQ they_PPHS2 appear_VV0 ,_, so_CS the_AT words_NN2 which_DDQ occur_VV0 in_II discourse_NN1 are_VBR constrained_VVN by_II what_DDQ ,_, following_VVG Halliday_NP1 ,_, we_PPIS2 shall_VM call_VVI their_APPGE co-text_NN1 ._. 
Consider_VV0 the_AT following_JJ lexical_JJ items_NN2 in_II a_AT1 number_NN1 of_IO verbal_JJ contexts_NN2 cited_VVD almost_RR at_RR21 random_RR22 from_II Darwin_NP1 's_GE Journal_NN1 during_II the_AT Voyage_NN1 of_IO HMS_NNB Beagle_NP1 round_II the_AT World_NN1 :_: (_( 16_MC )_) a_AT1 ._. 
The_AT children_NN2 of_IO the_AT Indians_NN2 are_VBR saved_VVN ,_, to_TO be_VBI sold_VVN or_CC given_VVN away_RL as_CSA servants_NN2 ,_, or_CC rather_RR slaves_NN2 for_IF as_RG long_RR a_AT1 time_NNT1 as_II the_AT owners_NN2 can_VM make_VVI them_PPHO2 believe_VVI themselves_PPX2 slaves_NN2 ._. 
But_CCB I_PPIS1 believe_VV0 in_II their_APPGE treatment_NN1 there_EX is_VBZ little_JJ to_TO complain_VVI of_IO (_( 114_MC )_) b_ZZ1 ._. 
The_AT same_DA evening_NNT1 I_PPIS1 went_VVD on_II shore_NN1 ._. 
The_AT first_MD landing_NN1 in_II any_DD new_JJ country_NN1 is_VBZ very_RG interesting._NNU (_( 169_MC )_) c_ZZ1 ._. 
When_CS we_PPIS2 came_VVD within_II hail_NN1 ,_, one_MC1 of_IO the_AT four_MC natives_NN2 who_PNQS were_VBDR present_JJ advanced_JJ to_TO receive_VVI us_PPIO2 and_CC began_VVD to_TO shout_VVI most_RGT vehemently_RR ,_, wishing_VVG to_TO direct_VVI us_PPIO2 where_RRQ to_TO land_VVI ._. 
When_CS we_PPIS2 were_VBDR on_II shore_NN1 the_AT party_NN1 looked_VVD rather_RG alarmed._NNU (_( 206_MC )_) d_ZZ1 ._. 
After_II crossing_VVG many_DA2 low_JJ hills_NN2 ,_, we_PPIS2 descended_VVD into_II the_AT small_JJ land-locked_JJ plain_NN1 of_IO Guitron_NN1 ._. 
In_II the_AT basins_NN2 ,_, such_II21 as_II22 this_DD1 one_PN1 ,_, which_DDQ are_VBR elevated_VVN from_II one_MC1 thousand_NNO to_II two_MC thousand_NNO feet_NN2 above_II the_AT sea_NN1 ,_, two_MC species_NN of_IO acacia_NN1 ..._... grow_VV0 in_II large_JJ numbers._NNU (_( 257_MC )_) (_( 1892_MC edition_NN1 )_) The_AT point_NN1 we_PPIS2 wish_VV0 to_TO make_VVI here_RL should_VM be_VBI an_AT1 obvious_JJ one_PN1 and_CC can_VM of_RR21 course_RR22 be_VBI made_VVN with_II31 respect_II32 to_II33 many_DA2 of_IO the_AT other_JJ items_NN2 which_DDQ we_PPIS2 have_VH0 not_XX italicised_VVN in_II the_AT cited_JJ texts_NN2 ._. 
However_RR ,_, consider_VV0 the_AT sort_NN1 of_IO lexical_JJ content_NN1 you_PPY would_VM expect_VVI to_TO find_VVI associated_VVN with_IW the_AT forms_NN2 treatment_NN1 ,_, landing_VVG ,_, party_NN1 and_CC basin_NN1 in_II a_AT1 dictionary_NN1 entry_NN1 ,_, and_CC note_VV0 how_RRQ finding_NN1 the_AT forms_NN2 embedded_VVN within_II a_AT1 co-text_NN1 constrains_VVZ their_APPGE interpretation_NN1 ._. 
Just_RR as_CSA the_AT interpretation_NN1 of_IO individual_JJ lexical_JJ items_NN2 is_VBZ constrained_VVN by_II co-text_NN1 ,_, so_RR is_VBZ the_AT interpretation_NN1 of_IO utterances_NN2 within_II a_AT1 discourse_NN1 ._. 
Consider_VV0 this_DD1 text_NN1 of_IO the_AT beginning_NN1 of_IO a_AT1 sixteenyear-old_JJ Scottish_JJ pupil_NN1 's_GE account_NN1 of_IO a_AT1 Semp_NN1 cartoon_NN1 :_: (_( 17_MC )_) a._NNU a_AT1 man_NN1 and_CC woman_NN1 sitting_VVG in_II the_AT living_NN1 room_NN1 +_FO the_AT woman_NN1 sitting_VVG reading_VVG quite_RG happily_RR the_AT man_NN1 's_GE bored_JJ goes_VVZ to_II the_AT window_NN1 looks_VVZ out_RP the_AT window_NN1 +_FO and_CC gets_VVZ himself_PPX1 ready_JJ and_CC goes_VVZ out_RP +_FO The_AT reader_NN1 must_VM interpret_VVI the_AT woman_NN1 sitting_VVG reading_VVG quite_RG happily_RR as_CSA the_AT '_GE woman_NN1 '_GE already_RR mentioned_VVN ,_, hence_RR must_VM construct_VVI an_AT1 interpretation_NN1 which_DDQ has_VHZ her_APPGE '_GE sitting_NN1 reading_VVG quite_RG happily_RR in_II the_AT living_NN1 room_NN1 '_GE ._. 
Similarly_RR the_AT window_NN1 which_DDQ the_AT man_NN1 approaches_NN2 must_VM be_VBI interpreted_VVN as_58 '_GE the_AT window_NN1 of_IO the_AT living_NN1 room_NN1 '_GE ._. 
The_AT speaker_NN1 continues_VVZ with_IW a_AT1 change_NN1 of_IO location_NN1 and_CC we_PPIS2 have_VH0 to_TO assume_VVI that_DD1 what_DDQ follows_VVZ is_VBZ within_II the_AT newly_RR introduced_VVN location_NN1 :_: b._NNU goes_VVZ to_II his_PPGE goes_VVZ to_II a_AT1 club_NN1 +_FO has_VHZ a_AT1 drink_NN1 talks_NN2 to_II the_AT barman_NN1 +_FO then_RT he_PPHS1 starts_VVZ dancing_VVG with_IW a_AT1 beautiful_JJ girl_NN1 long_RR black_JJ hair_NN1 +_FO has_VHZ a_AT1 good_JJ time_NNT1 +_FO We_PPIS2 interpret_VV0 everything_PN1 that_CST happens_VVZ here_RL as_CSA happening_VVG to_II the_AT man_NN1 we_PPIS2 met_VVD in_II the_AT living_NN1 room_NN1 who_PNQS is_VBZ now_RT at_II a_AT1 club_NN1 ._. 
So_RR he_PPHS1 has_VHZ a_AT1 drink_NN1 ,_, talks_NN2 to_II the_AT barman_NN1 ,_, starts_VVZ dancing_NN1 and_CC has_VHZ a_AT1 good_JJ time_NNT1 all_DB at_II the_AT '_GE club_NN1 '_GE ._. 
The_AT speaker_NN1 announces_VVZ another_DD1 change_NN1 of_IO location_NN1 c._ZZ1 then_RT he_PPHS1 goes_VVZ home_RL and_CC he_PPHS1 calls_VVZ her_APPGE +_FO and_CC his_APPGE wife_NN1 overhears_VVZ him_PPHO1 +_FO Again_RT we_PPIS2 assume_VV0 that_CST we_PPIS2 are_VBR still_RR talking_VVG about_II the_AT same_DA man_NN1 ,_, that_CST he_PPHS1 has_VHZ returned_VVN home_RL to_II the_AT location_NN1 where_CS the_AT '_GE living_NN1 room_NN1 '_GE we_PPIS2 first_MD met_VVD was_VBDZ located_VVN ._. 
Now_RT the_AT analyst_NN1 may_VM be_VBI in_II some_DD doubt_NN1 how_RRQ to_TO interpret_VVI and_CC he_PPHS1 calls_VVZ her_PPHO1 ,_, since_CS the_AT man_NN1 might_VM reasonably_RR go_VVI into_II the_AT house_NN1 and_CC call_NN1 (_( shout_VV0 for_IF )_) his_APPGE wife_NN1 ._. 
However_RRQV this_DD1 interpretation_NN1 is_VBZ ruled_VVN out_RP by_II the_AT following_JJ co-text_NN1 and_CC his_APPGE wife_NN1 overhears_VVZ him_PPHO1 ._. 
So_RR we_PPIS2 are_VBR obliged_VVN to_TO interpret_VVI calls_NN2 as_CSA meaning_NN1 '_GE phones_NN2 '_GE and_CC her_PPHO1 as_CSA referring_VVG to_II '_GE the_AT beautiful_JJ girl_NN1 with_IW long_JJ black_JJ hair_NN1 with_IW whom_PNQO he_PPHS1 danced_VVD and_CC had_VHD a_AT1 good_JJ time_NNT1 '_GE ._. 
Within_II the_AT co-text_NN1 ,_, as_CSA we_PPIS2 have_VH0 seen_VVN in_II (_( 17_MC )_) above_RL ,_, a_AT1 further_JJR context_NN1 may_VM be_VBI constructed_VVN which_DDQ has_VHZ its_APPGE own_DA index_NN1 of_IO coordinates_NN2 ._. 
Indeed_RR within_II that_DD1 constructed_JJ context_NN1 ,_, further_JJR contexts_NN2 may_VM be_VBI nested_VVN ._. 
Consider_VV0 the_AT following_JJ passages_NN2 :_: (_( 18_MC )_) About_RG four_MC months_NNT2 before_II the_AT time_NNT1 I_PPIS1 am_VBM writing_VVG of_IO ,_, my_APPGE Lady_NN1 had_VHD been_VBN in_II London_NP1 ,_, and_CC had_VHD gone_VVN over_II a_AT1 Reformatory_JJ ..._... 
The_AT matron_NN1 ,_, seeing_VVG my_APPGE Lady_NN1 took_VVD an_AT1 interest_NN1 in_II the_AT place_NN1 ,_, pointed_VVD out_RP a_AT1 girl_NN1 to_II her_PPHO1 ,_, named_VVN Rosanna_NP1 Spearman_NP1 ,_, and_CC told_VVN her_PPHO1 a_AT1 most_RGT miserable_JJ story_NN1 :_: which_DDQ I_PPIS1 have_VH0 n't_XX the_AT heart_NN1 to_TO repeat_VVI here_RL ;_; for_CS I_PPIS1 do_VD0 n't_XX like_VVI to_TO be_VBI made_VVN wretched_JJ without_IW any_DD use_NN1 ,_, and_CC no_AT more_DAR do_VD0 you_PPY ._. 
The_AT upshot_NN1 of_IO it_PPH1 was_VBDZ ,_, that_CST Rosanna_NP1 Spearman_NP1 had_VHD been_VBN a_AT1 thief_NN1 ..._... 
(_( Wilkie_NP1 Collins_NP1 ,_, The_AT Moonstone_NP1 )_) The_AT actual_JJ place_NN1 and_CC time_NNT1 of_IO writing_NN1 of_IO the_AT manuscript_NN1 by_II the_AT author_NN1 ,_, Wilkie_NP1 Collins_NP1 ,_, or_CC indeed_RR the_AT identity_NN1 of_IO the_AT author_NN1 ,_, is_VBZ not_XX a_AT1 necessary_JJ piece_NN1 of_IO information_NN1 for_IF the_AT reader_NN1 to_TO interpret_VVI the_AT text_NN1 ._. 
We_PPIS2 may_VM assume_VVI ,_, however_RR ,_, that_CST he_PPHS1 will_VM have_VHI a_AT1 better_JJR understanding_NN1 of_IO the_AT purpose_NN1 of_IO the_AT author_NN1 in_II constructing_VVG the_AT text_NN1 in_II the_AT way_NN1 it_PPH1 is_VBZ constructed_VVN if_CS he_PPHS1 knows_VVZ that_CST it_PPH1 is_VBZ written_VVN in_II the_AT late_JJ nineteenth_MD century_NNT1 (_( which_DDQ will_VM account_VVI for_IF some_DD differences_NN2 in_II code_NN1 ,_, in_II Hymes_NP1 '_GE terms_NN2 )_) in_II Victorian_JJ England_NP1 (_( which_DDQ will_VM account_VVI for_IF the_AT reference_NN1 to_II a_AT1 Reformatory_JJ )_) and_CC that_CST the_AT author_NN1 is_VBZ constructing_VVG the_AT first_MD English_JJ detective_NN1 story_NN1 ,_, narrating_VVG the_AT events_NN2 from_II the_AT point_NN1 of_IO view_NN1 of_IO four_MC different_JJ participants_NN2 ,_, whose_DDQGE characters_NN2 are_VBR in_RR21 part_RR22 revealed_VVN by_II the_AT narrative_JJ style_NN1 which_DDQ the_AT author_NN1 assigns_VVZ to_II them_PPHO2 ._. 
We_PPIS2 have_VH0 then_RT ,_, an_AT1 author_NN1 and_CC an_AT1 actual_JJ time_NNT1 and_CC place_NN1 of_IO writing_VVG the_AT novel_NN1 (_( or_CC a_AT1 series_NN of_IO times_NNT2 and_CC places_NN2 )_) ._. 
Then_RT to_II each_DD1 narrator_NN1 is_VBZ assigned_VVN a_AT1 time_NNT1 and_CC place_NN1 of_IO the_AT writing_NN1 of_IO his_APPGE contribution_NN1 ._. 
It_PPH1 is_VBZ presumably_RR that_DD1 time_NNT1 which_DDQ is_VBZ relevant_JJ to_II the_AT comment_NN1 which_DDQ I_PPIS1 have_VH0 n't_XX the_AT heart_NN1 to_TO repeat_VVI here_RL where_CS I_PPIS1 refers_VVZ to_II the_AT current_JJ narrator_NN1 ._. 
Immediately_RR preceding_VVG this_DD1 extracted_JJ fragment_NN1 ,_, the_AT narrator_NN1 has_VHZ been_VBN describing_VVG an_AT1 incident_NN1 relevant_JJ to_II the_AT main_JJ story_NN1 ._. 
This_DD1 is_VBZ referred_VVN to_II in_II the_AT expression_NN1 the_AT time_NNT1 I_PPIS1 am_VBM writing_VVG of_IO ._. 
The_AT narrator_NN1 then_RT proceeds_VVZ to_TO give_VVI some_DD background_NN1 information_NN1 ,_, which_DDQ he_PPHS1 situates_VVZ in_II a_AT1 previous_JJ time_NNT1 About_RG four_MC months_NNT2 before_RT ._. 
He_PPHS1 introduces_VVZ Rosanna_NP1 Spearman_NP1 ,_, who_PNQS ,_, at_II the_AT time_NNT1 four_MC months_NNT2 before_RT was_VBDZ a_AT1 resident_NN1 of_IO the_AT Reformatory_NP1 ,_, but_CCB at_II some_DD previous_JJ time_NNT1 to_II that_DD1 ,_, Rosanna_NP1 Spearman_NP1 had_VHD been_VBN a_AT1 thief_NN1 ._. 
Within_II the_AT time_NNT1 domain_NN1 of_IO '_GE four_MC months_NNT2 before_RT '_GE a_AT1 new_JJ speaker_NN1 and_CC hearer_NN1 are_VBR introduced_VVN :_: (_( 19_MC )_) My_APPGE Lady_NN1 ..._... said_VVD to_II the_AT matron_NN1 upon_II that_DD1 ,_, '_GE Rosanna_NP1 Spearman_NP1 shall_VM have_VHI her_APPGE chance_NN1 ,_, in_II my_APPGE service_NN1 '_GE ._. 
In_II a_AT1 week_NNT1 afterwards_RT ,_, Rosanna_NP1 Spearman_NP1 entered_VVD this_DD1 establishment_NN1 as_CSA our_APPGE second_MD housemaid_NN1 ._. 
At_II the_AT time_NNT1 of_IO utterance_NN1 ,_, four_MC months_NNT2 before_II the_AT time_NNT1 I_PPIS1 am_VBM writing_VVG of_IO ,_, the_AT beneficent_JJ lady_NN1 speaks_VVZ of_IO the_AT future_NN1 ,_, shall_VM have_VHI her_APPGE chance_NN1 ._. 
In_II the_AT following_JJ sentence_NN1 the_AT narrator_NN1 comments_NN2 on_II what_DDQ happened_VVD a_AT1 week_NNT1 later_RRR than_CSN the_AT time_NNT1 of_IO the_AT lady_NN1 's_GE speech_NN1 ,_, from_II the_AT point_NN1 of_IO view_NN1 of_IO his_APPGE context_NN1 at_II the_AT time_NNT1 of_IO writing_VVG his_APPGE contribution_NN1 to_II the_AT novel_NN1 ,_, In_II a_AT1 week_NNT1 afterwards_RT ..._... 
This_DD1 brief_JJ introduction_NN1 does_VDZ scant_VVI justice_NN1 to_II the_AT interest_NN1 of_IO the_AT temporal_JJ structure_NN1 of_IO this_DD1 passage_NN1 ._. 
It_PPH1 does_VDZ ,_, however_RR ,_, indicate_VVI the_AT complexity_NN1 of_IO nested_JJ contexts_NN2 established_VVN by_II co-text_NN1 which_DDQ ,_, as_CSA hearers_NN2 /_FO readers_NN2 ,_, we_PPIS2 are_VBR capable_JJ of_IO interpreting_VVG ._. 
In_II Chapter_NN1 6_MC we_PPIS2 shall_VM discuss_VVI the_AT issue_NN1 of_IO anaphoric_JJ reference_NN1 which_DDQ is_VBZ generally_RR held_VVN to_TO depend_VVI crucially_RR on_II co-text_NN1 for_IF interpretation_NN1 ._. 
For_IF the_AT moment_NN1 the_AT main_JJ point_NN1 we_PPIS2 are_VBR concerned_JJ to_TO make_VVI is_VBZ to_TO stress_VVI the_AT power_NN1 of_IO co-text_NN1 in_II constraining_VVG interpretation_NN1 ._. 
Even_RR in_II the_AT absence_NN1 of_IO information_NN1 about_II place_NN1 and_CC time_NNT1 of_IO original_JJ utterance_NN1 ,_, even_RR in_II the_AT absence_NN1 of_IO information_NN1 about_II the_AT speaker_NN1 /_FO writer_NN1 and_CC his_APPGE intended_JJ recipient_NN1 ,_, it_PPH1 is_VBZ often_RR possible_JJ to_TO reconstruct_VVI at_RR21 least_RR22 some_DD part_NN1 of_IO the_AT physical_JJ context_NN1 and_CC to_TO arrive_VVI at_II some_DD interpretation_NN1 of_IO the_AT text_NN1 ._. 
The_AT more_DAR co-text_NN1 there_EX is_VBZ ,_, in_RR21 general_RR22 ,_, the_AT more_RRR secure_VV0 the_AT interpretation_NN1 is_VBZ ._. 
Text_NN1 creates_VVZ its_APPGE own_DA context_NN1 ._. 
As_CSA Isard_NP1 (_( 1975_MC :_: 377_MC )_) remarks_VVZ :_: '_GE communications_NN2 do_VD0 not_XX merely_RR depend_VVI on_II the_AT context_NN1 for_IF their_APPGE interpretation_NN1 ,_, they_PPHS2 change_VV0 that_DD1 context_NN1 '_GE ._. 
The_AT expanding_JJ context_NN1 In_II our_APPGE discussion_NN1 so_RG far_RR ,_, we_PPIS2 have_VH0 been_VBN concerned_JJ to_TO impose_VVI some_DD sort_NN1 of_IO analytic_JJ structure_NN1 on_II the_AT lumpen_JJ mass_NN1 of_IO context_NN1 ._. 
We_PPIS2 have_VH0 abstracted_VVN away_II21 from_II22 particular_JJ contexts_NN2 ,_, across_II communicative_JJ contexts_NN2 in_RR21 general_RR22 ,_, to_TO arrive_VVI at_II a_AT1 set_NN1 of_IO features_NN2 ,_, some_DD of_IO which_DDQ seem_VV0 relevant_JJ to_II the_AT identification_NN1 of_IO a_AT1 speech_NN1 event_NN1 as_CSA being_VBG of_IO a_AT1 particular_JJ kind_NN1 ,_, to_II the_AT ability_NN1 of_IO the_AT hearer_NN1 to_TO predict_VVI what_DDQ sort_NN1 of_IO thing_NN1 the_AT speaker_NN1 is_VBZ likely_JJ to_TO say_VVI in_II a_AT1 given_JJ type_NN1 of_IO context_NN1 ,_, and_CC to_II the_AT constraining_NN1 of_IO interpretation_NN1 in_II context_NN1 ._. 
The_AT observant_JJ reader_NN1 will_VM have_VHI noticed_VVN that_CST we_PPIS2 have_VH0 helped_VVN ourselves_PPX2 to_II the_AT content_NN1 of_IO the_AT features_NN2 proposed_VVN by_II Hymes_NP1 and_CC the_AT co-ordinates_NN2 proposed_VVN by_II Lewis_NP1 in_II a_AT1 fairly_RR arbitrary_JJ way_NN1 ._. 
So_RR we_PPIS2 have_VH0 given_VVN variable_JJ amounts_NN2 of_IO information_NN1 about_II the_AT speaker_NN1 or_CC the_AT hearer_NN1 or_CC the_AT time_NNT1 or_CC the_AT place_NN1 as_CSA we_PPIS2 have_VH0 discussed_VVN different_JJ fragments_NN2 of_IO discourse_NN1 ._. 
This_DD1 behaviour_NN1 is_VBZ consistent_JJ with_IW Hymes_NP1 '_GE own_DA expectations_NN2 about_II how_RRQ his_APPGE framework_NN1 would_VM be_VBI used_VVN ._. 
You_PPY will_VM remember_VVI that_CST he_PPHS1 thought_VVD that_CST contextual_JJ features_NN2 might_VM be_VBI considered_VVN in_II the_AT way_NN1 that_CST general_JJ phonetic_JJ features_NN2 are_VBR considered_VVN :_: sometimes_RT ,_, but_CCB not_XX always_RR relevant_JJ ,_, and_CC specifiable_JJ to_II variable_JJ degrees_NN2 of_IO delicacy_NN1 for_IF different_JJ purposes_NN2 (_( 2.2_MC ._. 
I_ZZ1 )_) ._. 
A_AT1 problem_NN1 for_IF the_AT discourse_NN1 analyst_NN1 must_VM be_VBI ,_, then_RT ,_, to_TO decide_VVI when_RRQ a_AT1 particular_JJ feature_NN1 is_VBZ relevant_JJ to_II the_AT specification_NN1 of_IO a_AT1 particular_JJ context_NN1 and_CC what_DDQ degree_NN1 of_IO specification_NN1 is_VBZ required_VVN ._. 
Are_VBR there_RL general_JJ principles_NN2 which_DDQ will_VM determine_VVI the_AT relevance_NN1 or_CC nature_NN1 of_IO the_AT specification_NN1 ,_, or_CC does_VDZ the_AT analyst_NN1 have_VHI to_TO make_VVI ad_JJ21 hoc_JJ22 judgements_NN2 on_II these_DD2 questions_NN2 each_DD1 time_NNT1 he_PPHS1 attempts_VVZ to_TO work_VVI on_II a_AT1 fragment_NN1 of_IO discourse_NN1 ?_? 
For_IF the_AT moment_NN1 ,_, we_PPIS2 shall_VM limit_VVI our_APPGE discussion_NN1 of_IO this_DD1 question_NN1 to_II those_DD2 features_NN2 which_DDQ relate_VV0 directly_RR to_II the_AT deictic_JJ context_NN1 ,_, those_DD2 features_NN2 which_DDQ will_VM permit_VVI interpretation_NN1 for_IF deictic_JJ expressions_NN2 like_II the_AT temporal_JJ expression_NN1 now_RT ,_, the_AT spatial_JJ expression_NN1 here_RL ,_, and_CC the_AT first_MD person_NN1 expression_NN1 I._NP1 Are_VBR there_RL standard_JJ procedures_NN2 for_IF determining_VVG what_DDQ information_NN1 is_VBZ relevant_JJ to_II the_AT interpretation_NN1 of_IO these_DD2 expressions_NN2 ?_? 
Lyons_NP1 (_( 1977_MC :_: 570_MC )_) suggests_VVZ that_CST there_EX might_VM ,_, in_II principle_NN1 ,_, be_VBI such_DA standard_JJ procedures_NN2 :_: Every_AT1 actual_JJ utterance_NN1 is_VBZ spatiotemporally_RR unique_JJ ,_, being_VBG spoken_VVN or_CC written_VVN at_II a_AT1 particular_JJ place_NN1 and_CC at_II a_AT1 particular_JJ time_NNT1 ;_; and_CC provided_CS21 that_CS22 there_EX is_VBZ some_DD standard_JJ system_NN1 for_IF identifying_VVG points_NN2 in_II space_NN1 and_CC time_NNT1 ,_, we_PPIS2 can_VM ,_, in_II principle_NN1 ,_, specify_VV0 the_AT actual_JJ spatiotemporal_JJ situation_NN1 of_IO any_DD utterance_NN1 act_NN1 ._. 
There_RL clearly_RR are_VBR standard_JJ systems_NN2 for_IF locating_VVG points_NN2 in_II time_NNT1 and_CC space_NN1 ._. 
It_PPH1 would_VM be_VBI possible_JJ to_TO specify_VVI the_AT time_NNT1 of_IO an_AT1 utterance_NN1 as_CSA stretching_VVG between_RL say_VV0 9.33_MC a.m._RA and_CC 9.34_MC a.m._RA on_II 5_MC June_NPM1 1961_MC ,_, specifying_VVG the_AT utterance_NN1 in_II31 terms_II32 of_II33 clock_NN1 and_CC calendar_NN1 time_NNT1 ,_, good_JJ standard_JJ systems_NN2 ._. 
We_PPIS2 could_VM ,_, then_RT ,_, presumably_RR ,_, if_CS we_PPIS2 had_VHD the_AT relevant_JJ instrumentation_NN1 ,_, specify_VV0 the_AT place_NN1 of_IO the_AT utterance_NN1 in_II31 terms_II32 of_II33 a_AT1 fine_JJ interaction_NN1 of_IO latitude_NN1 and_CC longitude_NN1 ._. 
It_PPH1 is_VBZ not_XX at_RR21 all_RR22 clear_JJ ,_, however_RR ,_, that_CST these_DD2 particular_JJ standard_JJ systems_NN2 produce_VV0 the_AT relevant_JJ information_NN1 on_II all_DB occasions_NN2 ._. 
Presumably_RR some_DD patrol_NN1 ship_NN1 on_II the_AT high_JJ seas_NN2 might_VM log_VVI messages_NN2 in_II this_DD1 way_NN1 ,_, but_CCB it_PPH1 is_VBZ clear_JJ that_CST ,_, as_CSA humans_NN2 ,_, our_APPGE experience_NN1 of_IO utterances_NN2 is_VBZ not_XX that_CST we_PPIS2 have_VH0 recorded_VVN in_II memory_NN1 a_AT1 list_NN1 of_IO utterances_NN2 to_II which_DDQ are_VBR attached_VVN standard_JJ tags_NN2 specifying_VVG time_NNT1 and_CC place_VV0 in_II these_DD2 terms_NN2 ._. 
A_AT1 friend_NN1 can_VM attempt_VVI to_TO recall_VVI to_II your_APPGE mind_NN1 some_DD utterance_NN1 which_DDQ you_PPY both_DB2 experienced_VVN by_II a_AT1 variety_NN1 of_IO place_NN1 and_CC time_NNT1 tags_VVZ :_: (_( 20_MC )_) a_AT1 ._. 
But_CCB you_PPY just_RR said_VVD he_PPHS1 was_VBDZ n't_XX ._. 
(_( Place_VV0 :_: maintained_VVN ;_; time_NNT1 :_: only_JJ minutes_NNT2 ago_RA )_) b_ZZ1 ._. 
You_PPY said_VVD in_II the_AT staff_NN meeting_VVG yesterday_RT that_CST he_PPHS1 wasn't._NNU c_ZZ1 ._. 
You_PPY said_VVD last_MD week_NNT1 at_II the_AT staff_NN meeting_VVG that_CST he_PPHS1 wasn't._NNU d_ZZ1 ._. 
You_PPY said_VVD last_MD year_NNT1 when_RRQ we_PPIS2 met_VVD in_II Toronto_NP1 that_CST he_PPHS1 was_VBDZ n't_XX ._. 
The_AT further_RRR away_RL in_II time_NNT1 the_AT message_NN1 was_VBDZ situated_VVN ,_, the_AT less_RGR likely_RR the_AT speaker_NN1 is_VBZ to_TO remember_VVI precisely_RR the_AT date_NN1 and_CC time_NNT1 at_II which_DDQ it_PPH1 occurred_VVD ,_, and_CC the_AT larger_JJR the_AT time-span_NN1 he_PPHS1 is_VBZ likely_JJ to_TO make_VVI available_JJ for_IF it_PPH1 to_TO have_VHI occurred_VVN in_RP ._. 
It_PPH1 seems_VVZ unlikely_JJ then_RT ,_, that_DD1 '_VBZ standard_JJ procedures_NN2 '_GE of_IO recording_VVG space_NN1 and_CC time_NNT1 are_VBR going_VVGK to_TO be_VBI relevant_JJ to_II the_AT unique_JJ identification_NN1 of_IO utterance_NN1 acts_NN2 ._. 
Perhaps_RR the_AT standard_JJ procedures_NN2 will_VM enable_VVI us_PPIO2 to_TO fix_VVI the_AT relevant_JJ space_NN1 spans_VVZ for_IF the_AT interpretation_NN1 of_IO deictic_JJ expressions_NN2 like_II here_RL ._. 
Suppose_VV0 X_ZZ1 is_VBZ talking_VVG to_II Y_ZZ1 ,_, standing_VVG on_II the_AT blue_JJ border_NN1 of_IO the_AT carpet_NN1 in_II X_MC 's_GE office_NN1 ,_, in_II a_AT1 given_JJ street_NN1 ,_, in_II Manchester_NP1 ,_, in_II England_NP1 ,_, in_II Britain_NP1 ,_, in_II Western_NP1 Europe_NP1 ..._... 
Y_ZZ1 might_VM produce_VVI any_DD of_IO the_AT following_JJ utterances_NN2 :_: (_( 21_MC )_) a_AT1 ._. 
There_EX 's_VBZ another_DD1 worn_JJ section_NN1 which_DDQ needs_VVZ repair_NN1 here._NNU b_ZZ1 ._. 
You_PPY 've_VH0 got_VVN a_AT1 very_RG nice_JJ room_NN1 here._NNU c_ZZ1 ._. 
It_PPH1 's_VBZ a_AT1 really_RR nasty_JJ day_NNT1 here._NNU d_ZZ1 ._. 
You_PPY have_VH0 a_AT1 comparatively_RR mild_JJ climate_NN1 here_RL ._. 
It_PPH1 must_VM be_VBI clear_JJ that_CST the_AT spatial_JJ location_NN1 identified_VVN by_II here_RL in_II each_DD1 of_IO these_DD2 expressions_NN2 could_VM be_VBI interpreted_VVN as_II a_AT1 series_NN of_IO concentric_JJ rings_NN2 spreading_VVG out_RP from_II the_AT speaker_NN1 and_CC encompassing_VVG different_JJ amounts_NN2 of_IO physical_JJ space_NN1 ,_, but_CCB the_AT interpretation_NN1 of_IO the_AT spatial_JJ range_NN1 of_IO the_AT expression_NN1 here_RL on_II any_DD particular_JJ occasion_NN1 of_IO use_NN1 will_VM have_VHI to_TO be_VBI sought_VVN in_II the_AT context_NN1 of_IO what_DDQ the_AT speaker_NN1 is_VBZ talking_VVG about_II ._. 
What_DDQ appears_VVZ to_TO be_VBI stable_JJ in_II interpretations_NN2 of_IO here_RL (_( apart_II21 from_II22 curious_JJ usages_NN2 deriving_VVG from_II long-distance_JJ telephonic_JJ communication_NN1 and_CC long-distance_JJ travel_NN1 ,_, discussed_VVN in_II Lyons_NP1 ,_, 1977_MC )_) is_VBZ that_CST the_AT deictic_JJ centre_NN1 is_VBZ located_VVN where_CS the_AT speaker_NN1 is_VBZ ._. 
Very_RG similar_JJ problems_NN2 arise_VV0 with_IW the_AT interpretation_NN1 of_IO the_AT temporal_JJ deictic_JJ expression_NN1 now_RT ._. 
Consider_VV0 the_AT following_JJ possible_JJ utterances_NN2 :_: (_( 22_MC )_) a_AT1 ._. 
Clap_VV0 altogether_RR NOW._NP1 (_( gym_NN1 mistress_NN1 to_II class_NN1 )_) b_ZZ1 ._. 
I_PPIS1 think_VV0 you_PPY should_VM begin_VVI the_AT next_MD chapter_NN1 now._NNU (_( supervisor_NN1 to_II student_NN1 )_) c_ZZ1 ._. 
Now_RT I_PPIS1 'm_VBM getting_VVG older_JJR I_PPIS1 really_RR do_VD0 find_VVI policemen_NN2 look_VV0 younger._NNU d_ZZ1 ._. 
From_II the_AT iron_NN1 age_NN1 till_II now_RT ,_, man_NN1 has_VHZ been_VBN making_VVG increasingly_RR complex_JJ artefacts_NN2 ._. 
In_II c_ZZ1 and_CC d_ZZ1 the_AT utterances_NN2 appear_VV0 to_TO be_VBI located_VVN within_II different_JJ temporal_JJ spans_NN2 ,_, one_PN1 relating_VVG to_II the_AT speaker_NN1 's_GE advancing_JJ age_NN1 (_( involving_VVG a_AT1 span_NN1 of_IO 2030_MC years_NNT2 )_) as_II31 opposed_II32 to_II33 the_AT advancement_NN1 of_IO man_NN1 (_( involving_VVG a_AT1 span_NN1 at_RR21 least_RR22 of_IO decades_NNT2 and_CC possibly_RR centuries_NNT2 )_) ._. 
Utterances_NN2 a_AT1 and_CC b_ZZ1 are_VBR different_JJ in_CS21 that_CS22 the_AT action_NN1 specified_VVN is_VBZ to_TO follow_VVI the_AT utterance_NN1 ,_, immediately_RR in_II the_AT case_NN1 of_IO a_AT1 ,_, but_CCB after_CS some_DD expanse_NN1 of_IO time_NNT1 in_II b_ZZ1 ._. 
Once_RR21 again_RR22 we_PPIS2 suggest_VV0 that_CST the_AT deictic_JJ centre_NN1 is_VBZ located_VVN within_II the_AT context_NN1 of_IO utterance_NN1 by_II the_AT speaker_NN1 ,_, but_CCB that_CST the_AT interpretation_NN1 of_IO the_AT expression_NN1 now_RT as_CSA relating_VVG duratively_RR or_CC subsequently_RR to_II the_AT utterance_NN1 ,_, and_CC the_AT time-span_NN1 involved_VVD ,_, must_VM be_VBI determined_VVN with_II31 respect_II32 to_II33 the_AT content_NN1 of_IO the_AT utterance_NN1 ._. 
We_PPIS2 should_VM note_VVI that_CST this_DD1 fixing_NN1 of_IO the_AT deictic_JJ centre_NN1 is_VBZ particularly_RR appropriate_JJ to_II what_DDQ Lyons_NP1 (_( 1977_MC :_: 637_MC )_) calls_VVZ the_AT canonical_JJ situation_NN1 of_IO utterance_NN1 :_: this_DD1 involves_VVZ one-one_MC ,_, or_CC one-many_DA2 ,_, signalling_VVG in_II the_AT phonic_JJ medium_NN1 along_II the_AT vocal-auditory_JJ channel_NN1 ,_, with_IW all_DB the_AT participants_NN2 present_VV0 in_II the_AT same_DA actual_JJ situation_NN1 able_JK to_TO see_VVI one_PPX121 another_PPX122 and_CC to_TO perceive_VVI the_AT associated_JJ non-vocal_JJ paralinguistic_JJ features_NN2 of_IO their_APPGE utterances_NN2 ,_, and_CC each_DD1 assuming_VVG the_AT role_NN1 of_IO sender_NN1 and_CC receiver_NN1 in_II turn_NN1 ._. 
It_PPH1 is_VBZ ,_, of_RR21 course_RR22 ,_, possible_JJ to_TO use_VVI the_AT expressions_NN2 here_RL and_CC now_RT in_II what_DDQ might_VM be_VBI described_VVN as_II '_GE displaced_JJ contexts_NN2 '_GE ._. 
Consider_VV0 how_RRQ you_PPY would_VM interpret_VVI the_AT utterance_NN1 We_PPIS2 'll_VM land_VVI here_RL said_VVD by_II one_MC1 astronaut_NN1 to_II another_DD1 ,_, on_II earth_NN1 ,_, as_CSA they_PPHS2 study_VV0 a_AT1 map_NN1 of_IO the_AT moon_NN1 ._. 
Or_CC ,_, how_RRQ you_PPY interpret_VV0 the_AT message_NN1 on_II each_DD1 sheet_NN1 of_IO one_MC1 brand_NN1 of_IO government-issue_JJ toilet_NN1 roll_NN1 ,_, which_DDQ reads_VVZ NOW_RT WASH_VV0 YOUR_APPGE HANDS_NN2 ,_, PLEASE_RR ._. 
Speakers_NN2 ,_, or_CC writers_NN2 ,_, do_VD0 have_VHI the_AT option_NN1 of_IO transferring_VVG the_AT deictic_JJ centre_NN1 to_II the_AT hearer_NN1 's_GE ,_, or_CC reader_NN1 's_GE ,_, spatio-temporal_JJ situation_NN1 in_II which_DDQ the_AT text_NN1 will_VM be_VBI encountered_VVN ._. 
From_II our_APPGE discussion_NN1 of_IO the_AT spatio-temporal_JJ co-ordinates_NN2 which_DDQ seem_VV0 ,_, in_II principle_NN1 ,_, peculiarly_RR accessible_JJ to_II standard_JJ specification_NN1 ,_, it_PPH1 must_VM be_VBI obvious_JJ first_MD ,_, that_DD1 deictic_JJ expressions_NN2 may_VM retain_VVI a_AT1 standard_JJ deictic_JJ centre_NN1 but_CCB must_VM be_VBI interpreted_VVN with_II31 respect_II32 to_II33 the_AT content_NN1 of_IO the_AT utterance_NN1 in_II which_DDQ they_PPHS2 occur_VV0 and_CC ,_, second_NNT1 ,_, that_CST the_AT relevant_JJ standard_JJ temporal_JJ description_NN1 of_IO an_AT1 utterance_NN1 ,_, for_REX21 instance_REX22 9.22_MC a.m._RA on_II Tuesday_NPD1 28_MC June_NPM1 1873_MC ,_, as_II31 opposed_II32 to_II33 in_II the_AT late_JJ nineteenth_MD century_NNT1 ,_, will_VM vary_VVI depending_II21 on_II22 the_AT knowledge_NN1 and_CC intention_NN1 of_IO the_AT analyst_NN1 (_( or_CC speaker_NN1 )_) in_II referring_VVG to_II the_AT utterance_NN1 as_CSA located_VVN in_II time_NNT1 ._. 
That_REX41 is_REX42 to_REX43 say_REX44 ,_, even_CS21 if_CS22 there_EX were_VBDR an_AT1 agreed_JJ ,_, standard_JJ system_NN1 for_IF tagging_VVG utterances_NN2 with_IW spatio-temporal_JJ features_NN2 ,_, there_EX is_VBZ no_AT guarantee_NN1 that_CST that_DD1 tagging_JJ system_NN1 provides_VVZ the_AT relevant_JJ information_NN1 ._. 
Thus_RR in_II 2.2_MC ._. 
I._NP1 we_PPIS2 discussed_VVD a_AT1 fragment_NN1 of_IO discourse_NN1 :_: He_PPHS1 seemed_VVD to_TO resent_VVI them_PPHO2 on_II that_DD1 occasion_NN1 and_CC will_VM not_XX wear_VVI them_PPHO2 today_RT where_CS we_PPIS2 specified_VVD the_AT time_NNT1 of_IO utterance_NN1 as_CSA 4_MC June_NPM1 1980_MC ._. 
The_AT newspaper_NN1 article_NN1 from_II which_DDQ this_DD1 fragment_NN1 was_VBDZ extracted_VVN did_VDD indeed_RR appear_VVI on_II that_DD1 date_NN1 ._. 
However_RR ,_, for_IF anyone_PN1 who_PNQS knows_VVZ what_DDQ the_AT expression_NN1 the_AT Derby_NP1 means_VVZ ,_, it_PPH1 would_VM almost_RR certainly_RR have_VHI been_VBN more_RGR informative_JJ to_TO tag_VVI the_AT time_NNT1 of_IO utterance_NN1 as_CSA Derby_NP1 Day_NP1 ,_, 80_MC ._. 
The_AT space-time_JJ co-ordinates_NN2 can_VM not_XX be_VBI regarded_VVN as_CSA simple_JJ unstructured_JJ cues_NN2 to_II interpretation_NN1 in_II context_NN1 ._. 
Similarly_RR ,_, the_AT other_NN1 co-ordinates_VVZ relevant_JJ to_II the_AT deictic_JJ context_NN1 ,_, speaker_NN1 ,_, hearer_NN1 and_CC indicated_VVD object_NN1 ,_, can_VM not_XX be_VBI regarded_VVN as_CSA simple_JJ unstructured_JJ cues_NN2 which_DDQ demand_VV0 standard_JJ specification_NN1 ._. 
What_DDQ does_VDZ it_PPH1 mean_VVI to_TO specify_VVI ,_, for_REX21 instance_REX22 ,_, the_AT indicated_JJ object_NN1 co-ordinate_VV0 ?_? 
We_PPIS2 could_VM identify_VVI a_AT1 person_NN1 by_II name_NN1 ._. 
We_PPIS2 could_VM report_VVI Ellen_NP1 Blair_NP1 said_VVD she_PPHS1 'd_VM like_VVI to_TO come_VVI ._. 
This_DD1 might_VM be_VBI adequate_JJ to_TO identify_VVI the_AT speaker_NN1 ,_, indeed_RR the_AT expression_NN1 Ellen_NP1 might_VM be_VBI sufficient_JJ ._. 
If_CS ,_, however_RR ,_, you_PPY do_VD0 not_XX know_VVI who_PNQS this_DD1 person_NN1 is_VBZ ,_, or_CC might_VM be_VBI ,_, it_PPH1 would_VM be_VBI more_RGR helpful_JJ if_CS we_PPIS2 were_VBDR to_TO give_VVI some_DD indication_NN1 of_IO why_RRQ we_PPIS2 have_VH0 introduced_VVN her_PPHO1 into_II the_AT conversation_NN1 ._. 
So_RR we_PPIS2 might_VM say_VVI my_APPGE friend_NN1 Ellen_NP1 Blair_NP1 ,_, or_CC the_AT former_DA chairman_NN1 Ellen_NP1 Blair_NP1 ,_, or_CC a_AT1 nurse_NN1 in_II the_AT ward_NN1 called_VVN Ellen_NP1 Blair_NP1 ,_, giving_VVG ,_, in_II some_DD sense_NN1 ,_, '_GE credentials_NN2 '_GE for_IF her_APPGE existence_NN1 and_CC for_IF her_APPGE relationship_NN1 to_II the_AT speaker_NN1 who_PNQS is_VBZ responsible_JJ for_IF introducing_VVG her_PPHO1 into_II the_AT conversation_NN1 ._. 
Morgan_NP1 (_( 1975_MC :_: 442_MC )_) asks_58 '_GE What_DDQ can_VM we_PPIS2 infer_VVI about_II the_AT speaker_NN1 's_GE intentions_NN2 from_II the_AT fact_NN1 that_CST he_PPHS1 has_VHZ chosen_VVN this_DD1 particular_JJ description_NN1 ,_, rather_II21 than_II22 any_DD of_IO the_AT others_NN2 which_DDQ would_VM call_VVI to_TO mind_VVI the_AT same_DA referent_NN1 ?_? 
'_" For_IF any_DD individual_NN1 there_EX will_VM be_VBI an_AT1 immense_JJ number_NN1 of_IO possible_JJ descriptions_NN2 which_DDQ will_VM be_VBI more_RGR or_CC less_RGR appropriate_JJ in_II different_JJ contexts_NN2 ._. 
We_PPIS2 may_VM identify_VVI the_AT person_NN1 from_II external_JJ physical_JJ cues_NN2 :_: the_AT woman_NN1 in_II the_AT corner_NN1 ,_, the_AT man_NN1 with_IW a_AT1 beard_NN1 ,_, the_AT student_NN1 who_PNQS has_VHZ had_VHN his_APPGE hair_NN1 dyed_VVN ,_, the_AT child_NN1 in_II the_AT pink_JJ dress_NN1 or_CC ,_, more_RGR or_CC less_RGR flatteringly_RR ,_, the_AT tall_JJ distinguished_JJ looking_JJ man_NN1 I_PPIS1 the_AT man_NN1 with_IW a_AT1 big_JJ nose_NN1 and_CC stringy_JJ hair_NN1 ._. 
We_PPIS2 may_VM identify_VVI people_NN from_II a_AT1 description_NN1 of_IO what_DDQ they_PPHS2 are_VBR doing_VDG :_: the_AT woman_NN1 who_PNQS is_VBZ chatting_VVG up_RP the_AT Admiral_NN1 ,_, the_AT man_NN1 who_PNQS 's_VBZ fixing_VVG the_AT car_NN1 ,_, etc_RA ._. 
The_AT variable_NN1 which_DDQ interests_VVZ us_PPIO2 most_DAT is_VBZ that_DD1 which_DDQ is_VBZ concerned_JJ with_IW the_AT various_JJ roles_NN2 played_VVN by_II the_AT individual_NN1 ._. 
Lyons_NP1 (_( 1977_MC :_: 574ff._FO )_) distinguishes_VVZ between_II the_AT deictic_JJ role_NN1 of_IO an_AT1 individual_NN1 (_( which_DDQ assigns_VVZ ,_, for_REX21 instance_REX22 ,_, first_MD ,_, second_MD and_CC third_MD person_NN1 pronouns_NN2 )_) and_CC his_APPGE social_JJ role_NN1 or_CC '_GE status_NN1 '_GE ._. 
Lyons_NP1 points_VVZ out_RP that_DD1 ,_, for_REX21 example_REX22 ,_, the_AT terms_NN2 of_IO address_NN1 used_VVN by_II a_AT1 social_JJ inferior_JJ to_II a_AT1 social_JJ superior_NN1 may_VM be_VBI different_JJ from_II those_DD2 used_VVN between_II peers_NN2 ,_, as_CSA in_II vocative_JJ terms_NN2 like_NN1 '_GE Sir_NN1 '_GE or_CC '_GE Doctor_NN1 '_GE or_CC '_" My_APPGE Lord_NN1 '_GE (_( in_II the_AT courtroom_NN1 )_) ._. 
In_II different_JJ social_JJ contexts_NN2 ,_, then_RT ,_, different_JJ terms_NN2 of_IO address_NN1 will_VM be_VBI found_VVN ._. 
(_( Consider_VV0 for_REX21 instance_REX22 ,_, the_AT distribution_NN1 of_IO the_AT tu_NN1 /_FO vous_FW pronouns_NN2 in_II French_NN1 ._. )_) 
In_RR21 general_RR22 we_PPIS2 may_VM assume_VVI that_DD1 ,_, in_II a_AT1 particular_JJ social_JJ context_NN1 ,_, only_RR one_MC1 role_NN1 is_VBZ taken_VVN by_II an_AT1 individual_NN1 at_II one_MC1 particular_JJ time_NNT1 ._. 
A_AT1 glance_NN1 at_II any_DD newspaper_NN1 will_VM yield_VVI a_AT1 rich_JJ crop_NN1 of_IO identifications_NN2 of_IO individuals_NN2 in_II31 terms_II32 of_II33 the_AT social_JJ role_NN1 relevant_JJ to_II the_AT news_NN1 item_NN1 ._. 
Here_RL are_VBR just_RR a_AT1 few_DA2 :_: (_( 23_MC )_) a_AT1 ._. 
Daily_JJ Telegraph_NN1 cartoonist_NN1 Nicholas_NP1 Carland_NP1 showing_VVG how_RRQ he_PPHS1 sees_VVZ the_AT Prime_JJ Minister_NN1 ._. 
(_( Stop_VV0 Press_NN1 ,_, 27_MC February_NPM1 1982_MC )_) b_ZZ1 ._. 
Frank_NP1 Silbey_NP1 ,_, chief_JJ investigator_NN1 for_IF the_AT Senate_NN1 Labor_NN1 and_CC Human_JJ Resources_NN2 Committee_NN1 ,_, picked_VVD up_RP his_APPGE telephone_NN1 ._. 
(_( Time_NNT1 ,_, 31_MC May_NPM1 1982_MC )_) c_ZZ1 ._. 
Sophia_NP1 Loren_NP1 ,_, the_AT film_NN1 actress_NN1 ,_, awoke_VVD in_II a_AT1 prison_NN1 cell_NN1 in_II Caserta_NP1 ,_, near_II Naples_NP1 ,_, today_RT ._. 
(_( The_AT Times_NNT2 ,_, 21_MC May_NPM1 1982d._NNU d_ZZ1 ._. 
Mr._NNB Robert_NP1 Mugabe_NP1 ,_, the_AT Prime_JJ Minister_NN1 of_IO Zimbabwe_NP1 yesterday_RT sought_VVD to_TO reassure_VVI prospective_JJ investors_NN2 in_II his_APPGE country_NN1 ._. 
(_( The_AT Times_NNT2 ,_, 21_MC May_NPM1 1982_MC )_) e_ZZ1 ._. 
Senor_NNB Jorge_NP1 Blanco_NP1 of_IO the_AT ruling_JJ Revolutionary_JJ Party_NN1 was_VBDZ officially_RR declared_VVN winner_NN1 ._. 
(_( The_AT Times_NNT2 ,_, 21_MC May_NPM1 1982_MC )_) In_II each_DD1 case_NN1 the_AT individual_NN1 is_VBZ identified_VVN either_RR by_II the_AT role_NN1 which_DDQ is_VBZ relevant_JJ to_II the_AT content_NN1 of_IO the_AT article_NN1 ,_, or_CC by_II the_AT role_NN1 by_II which_DDQ he_PPHS1 is_VBZ known_VVN to_II the_AT public_NN1 ._. 
Each_DD1 individual_NN1 may_VM play_VVI many_DA2 other_JJ roles_NN2 parent_NN1 ,_, child_NN1 ,_, niece_NN1 ,_, brother_NN1 ,_, chess_NN1 player_NN1 ,_, gardener_NN1 ,_, diarist_NN1 ,_, but_CCB these_DD2 roles_NN2 are_VBR not_XX relevant_JJ in_II this_DD1 context_NN1 ,_, so_RR not_XX mentioned_VVN on_II this_DD1 occasion_NN1 ._. 
It_PPH1 is_VBZ possible_JJ for_IF more_DAR than_CSN one_MC1 social_JJ role_NN1 to_TO be_VBI relevant_JJ at_II one_MC1 time_NNT1 ._. 
Rommetveit_NP1 (_( 1974_MC :_: 45_MC )_) discusses_VVZ a_AT1 sentence_NN1 introduced_VVN in_II Chomsky_NP1 (_( 1972_MC :_: 67_MC )_) :_: I_PPIS1 am_VBM not_XX against_II MY_APPGE FATHER_NN1 ,_, only_RR against_II THE_AT LABOR_NN1 MINISTER_NN1 Rommetveit_NN1 argues_VVZ that_CST the_AT sentence_NN1 is_VBZ not_XX necessarily_RR self-contradictory_JJ even_CS21 if_CS22 the_AT individual_NN1 referred_VVN to_II by_II the_AT two_MC nominal_JJ expressions_NN2 is_VBZ the_AT same_DA individual_NN1 ._. 
It_PPH1 merely_RR expresses_VVZ the_AT ambivalence_NN1 which_DDQ is_VBZ a_AT1 common_JJ human_JJ experience_NN1 where_CS some_DD aspect_NN1 of_IO an_AT1 entity_NN1 pleases_VVZ you_PPY and_CC some_DD other_JJ aspect_NN1 fails_VVZ to_TO please_VVI ._. 
Rommetveit_NN1 argues_VVZ against_II '_GE the_AT notion_NN1 of_IO identifying_VVG reference_NN1 as_II an_AT1 unequivocally_RR defined_VVN point_NN1 in_II a_AT1 monistic_JJ and_CC epistemological_JJ transparent_JJ space_NN1 ,_, constructed_VVN on_II axiomatic_JJ prerequisites_NN2 for_IF specific_JJ operations_NN2 within_II formal_JJ logic_NN1 '_GE ..._... where_CS '_GE the_AT severe_JJ laws_NN2 of_IO truth_NN1 values_NN2 prescribe_VV0 that_CST the_AT speaker_NN1 must_VM know_VVI him_PPHO1 (_( the_AT indicated_JJ entity_NN1 )_) fully_RR or_CC not_XX at_RR21 all_RR22 '_GE (_( 1974_MC :_: 48_MC )_) ._. 
It_PPH1 is_VBZ possible_JJ for_IF speakers_NN2 ,_, hearers_NN2 or_CC indicated_VVD entities_NN2 to_TO be_VBI regarded_VVN from_II the_AT perspective_NN1 of_IO more_DAR than_CSN one_MC1 role_NN1 ._. 
Consider_VV0 :_: (_( 24_MC )_) a_AT1 ._. 
As_CSA his_APPGE neighbour_NN1 I_PPIS1 see_VV0 quite_RG a_AT1 lot_NN1 of_IO him_PPHO1 ,_, as_CSA his_APPGE colleague_NN1 I_PPIS1 hardly_RR ever_RR see_VV0 him._NNU b_ZZ1 ._. 
As_II a_AT1 colleague_NN1 you_PPY 're_VBR deficient_JJ but_CCB as_II a_AT1 neighbour_NN1 you_PPY 're_VBR marvellous._NNU c_ZZ1 ._. 
I_PPIS1 quite_RR like_VV0 her_PPHO1 as_II a_AT1 colleague_NN1 and_CC she_PPHS1 's_VBZ very_RG pleasant_JJ as_II a_AT1 casual_JJ friend_NN1 but_CCB she_PPHS1 's_VBZ impossible_JJ to_TO live_VVI with_IW ._. 
It_PPH1 is_VBZ clear_JJ that_CST we_PPIS2 can_VM hold_VVI partially_RR or_CC severely_RR differing_JJ opinions_NN2 about_II the_AT same_DA individual_NN1 in_II different_JJ roles_NN2 ._. 
In_II the_AT following_JJ extract_NN1 from_II a_AT1 report_NN1 in_II The_AT Times_NNT2 (_( 15_MC May_NPM1 1982_MC )_) the_AT same_DA individual_NN1 is_VBZ referred_VVN to_II by_II a_AT1 number_NN1 of_IO different_JJ expressions_NN2 which_DDQ relate_VV0 to_II the_AT multiple_JJ roles_NN2 that_CST the_AT reporter_NN1 considers_VVZ relevant_JJ to_II the_AT incident_NN1 :_: (_( 25_MC )_) Priest_NN1 is_VBZ charged_VVN with_IW Pope_NP1 attack_NN1 (_( Lisbon_NP1 ,_, May_NPM1 14_MC )_) A_ZZ1 dissident_JJ Spanish_JJ priest_NN1 was_VBDZ charged_VVN here_RL today_RT with_IW attempting_VVG to_TO murder_VVI the_AT Pope_NN1 ._. 
Juan_NP1 Fernandez_NP1 Krohn_NP1 ,_, aged_II 32_MC ,_, was_VBDZ arrested_VVN after_II a_AT1 man_NN1 armed_VVN with_IW a_AT1 bayonet_NN1 approached_VVD the_AT Pope_NN1 while_CS he_PPHS1 was_VBDZ saying_VVG prayers_NN2 at_II Fatima_NP1 on_II Wednesday_NPD1 night_NNT1 ._. 
According_II21 to_II22 the_AT police_NN2 ,_, Fernandez_NP1 told_VVD the_AT investigating_JJ magistrates_NN2 today_RT he_PPHS1 had_VHD trained_VVN for_IF the_AT past_JJ six_MC months_NNT2 for_IF the_AT assault_NN1 ._. 
He_PPHS1 was_VBDZ alleged_VVN to_TO have_VHI claimed_VVN the_AT Pope_NP1 '_GE looked_VVD furious_58 '_GE on_II hearing_VVG the_AT priest_NN1 's_GE criticism_NN1 of_IO his_APPGE handling_NN1 of_IO the_AT church_NN1 's_GE affairs_NN2 ._. 
If_CS found_VVN guilty_JJ ,_, the_AT Spaniard_NN1 faces_VVZ a_AT1 prison_NN1 sentence_NN1 of_IO 1520_MC years_NNT2 ._. 
We_PPIS2 have_VH0 italicised_VVN the_AT expressions_NN2 relating_VVG to_II the_AT man_NN1 identified_VVN in_II the_AT headline_NN1 as_CSA Priest_NN1 ._. 
The_AT relevance_NN1 of_IO his_APPGE role_NN1 as_CSA priest_NN1 (_( referred_VVN to_II by_II the_AT expressions_NN2 Priest_NN1 ,_, a_AT1 dissident_NN1 ..._... priest_NN1 ,_, the_AT priest_NN1 's_GE )_) is_VBZ presumably_RR as_II a_AT1 priest_NN1 of_IO the_AT Roman_JJ Catholic_JJ Church_NN1 of_IO which_DDQ the_AT Pope_NN1 is_VBZ Head_NN1 ._. 
Since_CS the_AT incident_NN1 reported_VVN takes_VVZ place_NN1 in_II Portugal_NP1 (_( Lisbon_NP1 )_) and_CC any_DD subsequent_JJ prison_NN1 sentence_NN1 will_VM be_VBI served_VVN in_II Portugal_NP1 ,_, it_PPH1 is_VBZ relevant_JJ that_CST the_AT priest_NN1 is_VBZ not_XX Portuguese_JJ (_( a_AT1 ..._... 
Spanish_JJ priest_NN1 ,_, the_AT Spaniard_NN1 )_) ._. 
A_AT1 potentially_RR confusing_JJ indefinite_JJ referring_VVG expression_NN1 ,_, a_AT1 man_NN1 armed_VVN with_IW a_AT1 bayonet_NN1 ,_, apparently_RR relates_VVZ back_RP to_II the_AT period_NN1 before_CS he_PPHS1 was_VBDZ identified_VVN as_58 '_GE a_AT1 dissident_JJ Spanish_JJ priest_NN1 '_GE ._. 
He_PPHS1 is_VBZ identified_VVN by_II his_APPGE name_NN1 ,_, as_CSA an_AT1 individual_JJ ,_, in_II the_AT set_NN1 constituted_VVN by_II the_AT intersection_NN1 of_IO the_AT various_JJ relevant_JJ roles_NN2 (_( Juan_NP1 Fernandez_NP1 Krohn_NP1 ,_, Fernandez_NP1 )_) ._. 
As_CSA Levy_NP1 (_( 1979_MC :_: 193_MC )_) remarks_VVZ ,_, '_GE the_AT speaker_NN1 by_II making_VVG reference_NN1 may_VM not_XX simply_RR identify_VVI but_CCB may_VM construct_VVI the_AT object_NN1 by_II selecting_VVG from_II a_AT1 field_NN1 of_IO relations_NN2 those_DD2 properties_NN2 that_CST are_VBR relevant_JJ at_II the_AT moment_NN1 of_IO utterance_NN1 '_GE ._. 
Consider_VV0 the_AT response_NN1 of_IO a_AT1 five-and-a-half-year-old_JJ girl_NN1 in_II a_AT1 Yorkshire_NP1 infant_NN1 school_NN1 where_CS she_PPHS1 is_VBZ asked_VVN to_TO say_VVI how_RRQ two_MC pictures_NN2 are_VBR different_JJ from_II each_PPX221 other_PPX222 ._. 
She_PPHS1 replies_VVZ :_: (_( 26_MC )_) a_AT1 ._. 
That_DD1 one_PN1 's_VBZ over_RP there_RL in_II that_DD1 but_CCB it_PPH1 i_MC1 n't_XX there_RL ._. 
The_AT teacher_NN1 then_RT holds_VVZ the_AT little_JJ girl_NN1 's_GE hands_NN2 ,_, so_CS she_PPHS1 ca_VM n't_XX point_VVI ,_, shuts_VVZ her_APPGE own_DA eyes_NN2 and_CC says_VVZ to_II the_AT child_NN1 :_: b_ZZ1 ._. 
Now_RT I_PPIS1 ca_VM n't_XX see_VVI the_AT picture_NN1 ._. 
Tell_VV0 me_PPIO1 the_AT difference_NN1 again_RT ._. 
This_DD1 time_NNT1 the_AT child_NN1 says_VVZ :_: c_ZZ1 ._. 
In_II this_DD1 picture_NN1 the_AT teddy_NN1 's_VBZ on_II the_AT chair_NN1 but_CCB there_EX ai_FU n't_XX no_AT teddy_NN1 in_II that_DD1 one_PN1 ._. 
The_AT pictures_NN2 are_VBR identical_JJ except_CS in_II three_MC respects_NN2 :_: the_AT presence_NN1 or_CC absence_NN1 of_IO a_AT1 teddy_NN1 bear_NN1 sitting_VVG on_II the_AT chair_NN1 ,_, a_AT1 difference_NN1 in_II the_AT pattern_NN1 on_II the_AT counterpane_NN1 ,_, a_AT1 difference_NN1 in_II the_AT position_NN1 of_IO a_AT1 mirror_NN1 ._. 
For_IF the_AT child_NN1 the_AT teddy_NN1 bear_NN1 is_VBZ clearly_RR the_AT salient_JJ object_NN1 ._. 
She_PPHS1 relies_VVZ in_II her_APPGE first_MD response_NN1 on_II the_AT teacher_NN1 's_GE access_NN1 to_II the_AT shared_JJ visual_JJ context_NN1 to_TO interpret_VVI what_DDQ she_PPHS1 says_VVZ ._. 
She_PPHS1 points_VVZ to_II the_AT teddy_NN1 bear_NN1 (_( that_DD1 one_MC1 )_) in_II the_AT first_MD picture_NN1 and_CC then_RT points_VVZ to_II the_AT empty_JJ chair_NN1 in_II the_AT second_MD picture_NN1 (_( there_RL )_) and_CC assumes_VVZ that_CST the_AT teacher_NN1 is_VBZ paying_VVG attention_NN1 to_II what_DDQ she_PPHS1 is_VBZ pointing_VVG to_II in_II their_APPGE shared_JJ context_NN1 of_IO situation_NN1 ._. 
When_CS the_AT teacher_NN1 inhibits_VVZ the_AT child_NN1 from_II pointing_VVG and_CC pretends_VVZ not_XX to_TO be_VBI able_JK to_TO see_VVI the_AT picture_NN1 ,_, the_AT child_NN1 understands_VVZ that_CST the_AT communicative_JJ situation_NN1 has_VHZ changed_VVN ,_, that_CST she_PPHS1 can_VM no_RR21 longer_RR22 rely_VVI on_II the_AT shared_JJ visual_JJ context_NN1 and_CC she_PPHS1 makes_VVZ her_APPGE reference_NN1 explicit_JJ (_( the_AT teddy_NN1 )_) ,_, locates_VVZ him_PPHO1 verbally_RR rather_CS21 than_CS22 by_II pointing_VVG to_II him_PPHO1 (_( on_II the_AT chair_NN1 )_) and_CC makes_VVZ explicit_JJ how_RRQ the_AT second_MD picture_NN1 differs_VVZ from_II the_AT first_MD (_( there_RL ai_FU n't_XX no_AT teddy_NN1 )_) ._. 
A_AT1 salient_JJ aspect_NN1 of_IO the_AT addressee_NN1 ,_, her_APPGE ability_NN1 to_TO see_VVI what_DDQ the_AT child_NN1 can_VM see_VVI ,_, has_VHZ been_VBN changed_VVN by_II the_AT utterance_NN1 of_IO b_ZZ1 and_CC the_AT acts_NN2 accompanying_VVG the_AT utterance_NN1 ._. 
Speakers_NN2 ,_, hearers_NN2 and_CC indicated_VVD objects_NN2 are_VBR not_XX featureless_JJ ,_, colourless_JJ spheres_NN2 ._. 
Nor_CC do_VD0 they_PPHS2 come_VVI simply_RR tagged_VVN with_IW proper_JJ names_NN2 appropriate_VV0 to_II all_DB occasions_NN2 together_RL with_IW one_PN1 identifying_VVG description_NN1 appropriate_JJ to_II all_DB occasions_NN2 ._. 
They_PPHS2 are_VBR ,_, characteristically_RR ,_, endowed_VVN with_IW immense_JJ numbers_NN2 of_IO physical_JJ and_CC social_JJ properties_NN2 ,_, any_DD one_MC1 of_IO which_DDQ may_VM be_VBI the_AT property_NN1 which_DDQ is_VBZ relevant_JJ to_II a_AT1 particular_JJ communicative_JJ act_NN1 ._. 
The_AT philosopher_NN1 's_GE crisp_JJ index_NN1 ,_, which_DDQ permits_VVZ the_AT identification_NN1 of_IO speaker_NN1 and_CC hearer_NN1 as_CSA X_ZZ1 and_CC Y_ZZ1 ,_, is_VBZ only_RR relevant_JJ in_II a_AT1 restricted_JJ model_NN1 world_NN1 ._. 
The_AT discourse_NN1 analyst_NN1 working_VVG in_II the_AT real_JJ world_NN1 has_VHZ to_TO be_VBI able_JK to_TO extract_VVI ,_, see_VV0 as_RG relevant_JJ ,_, just_RR those_DD2 properties_NN2 of_IO the_AT features_NN2 of_IO context_NN1 which_DDQ are_VBR relevant_JJ to_II the_AT particular_JJ communicative_JJ act_NN1 which_DDQ he_PPHS1 is_VBZ describing_VVG ,_, and_CC which_DDQ contribute_VV0 to_II the_AT interpretation_NN1 (_( or_CC intended_JJ meaning_NN1 )_) of_IO the_AT utterance_NN1 ._. 
As_CSA Enkvist_JJ (_( 1980_MC :_: 79_MC )_) remarks_VVZ ,_, '_GE The_AT context_NN1 analyst_NN1 's_GE first_MD embarrassment_NN1 is_VBZ richness_NN1 ._. 
'_" How_RRQ is_VBZ he_PPHS1 to_TO determine_VVI which_DDQ properties_NN2 of_IO which_DDQ features_NN2 of_IO context_NN1 are_VBR relevant_JJ on_II a_AT1 particular_JJ occasion_NN1 ?_? 
Are_VBR there_RL general_JJ principles_NN2 to_TO appeal_VVI to_II ?_? 
Is_VBZ it_PPH1 reasonable_JJ to_TO assume_VVI ,_, as_CSA we_PPIS2 tend_VV0 to_TO do_VDI ,_, that_CST those_DD2 features_NN2 of_IO context_NN1 which_DDQ are_VBR salient_JJ to_II the_AT speaker_NN1 are_VBR equally_RR salient_JJ to_II the_AT hearer_NN1 ?_? 
Ought_VMK we_PPIS2 not_XX rather_RR to_TO think_VVI in_II31 terms_II32 of_II33 partially_RR intersecting_VVG views_NN2 of_IO context_NN1 ?_? 
Bar-Hillel_NP1 (_( 1970_MC :_: 79_MC )_) states_VVZ that_DD1 '_VBZ the_AT depth_NN1 of_IO the_AT pragmatic_JJ context_NN1 which_DDQ is_VBZ necessary_JJ for_IF the_AT full_JJ understanding_NN1 of_IO various_JJ sentence-tokens_NN2 ,_, is_VBZ different_JJ ,_, of_RR21 course_RR22 ,_, from_II case_NN1 to_II case_NN1 '_GE ._. 
As_RR21 yet_RR22 we_PPIS2 have_VH0 only_RR a_AT1 very_RG limited_JJ understanding_NN1 of_IO how_RRQ we_PPIS2 might_VM set_VVI about_RP determining_JJ '_GE the_AT depth_NN1 of_IO the_AT pragmatic_JJ context_NN1 which_DDQ is_VBZ necessary_JJ '_GE for_IF interpretation_NN1 ._. 
We_PPIS2 outline_VV0 a_AT1 possible_JJ approach_NN1 to_II the_AT problem_NN1 in_II the_AT next_MD section_NN1 and_CC in_II Chapter_NN1 3_MC ._. 
The_AT principles_NN2 of_IO '_GE local_JJ interpretation_NN1 '_GE and_CC of'_II analogy'_NN1 In_II 2.3_MC we_PPIS2 have_VH0 discussed_VVN the_AT problems_NN2 for_IF the_AT discourse_NN1 analyst_NN1 in_II specifying_VVG what_DDQ aspects_NN2 of_IO the_AT apparently_RR illimitable_JJ features_NN2 of_IO context_NN1 are_VBR to_TO be_VBI taken_VVN into_II account_NN1 in_II the_AT interpretation_NN1 of_IO discourse_NN1 ._. 
How_RRQ is_VBZ he_PPHS1 to_TO determine_VVI the_AT relevant_JJ span_NN1 of_IO time_NNT1 in_II the_AT interpretation_NN1 of_IO a_AT1 particular_JJ utterance_NN1 of_IO '_" now_RT '_GE or_CC the_AT relevant_JJ aspects_NN2 of_IO a_AT1 character_NN1 referred_VVN to_II by_II the_AT expression_NN1 '_GE John_NP1 '_GE ?_? 
We_PPIS2 must_VM assume_VVI that_CST the_AT problem_NN1 for_IF the_AT discourse_NN1 analyst_NN1 is_VBZ ,_, in_II this_DD1 case_NN1 ,_, identical_JJ to_II the_AT problem_NN1 for_IF the_AT hearer_NN1 ._. 
There_EX must_VM be_VBI principles_NN2 of_IO interpretation_NN1 available_JJ to_II the_AT hearer_NN1 which_DDQ enable_VV0 him_PPHO1 to_TO determine_VVI ,_, for_REX21 instance_REX22 ,_, a_AT1 relevant_JJ and_CC reasonable_JJ interpretation_NN1 of_IO an_AT1 expression_NN1 '_GE John_NP1 '_GE on_II a_AT1 particular_JJ occasion_NN1 of_IO utterance_NN1 ._. 
One_MC1 principle_NN1 which_DDQ we_PPIS2 can_VM identify_VVI we_PPIS2 shall_VM call_VVI the_AT principle_NN1 of_IO local_JJ interpretation_NN1 ._. 
This_DD1 principle_NN1 instructs_VVZ the_AT hearer_NN1 not_XX to_TO construct_VVI a_AT1 context_NN1 any_DD larger_JJR than_CSN he_PPHS1 needs_VVZ to_TO arrive_VVI at_II an_AT1 interpretation_NN1 ._. 
Thus_RR if_CS he_PPHS1 hears_VVZ someone_PN1 say_VVI '_GE Shut_VV0 the_AT door_NN1 '_GE he_PPHS1 will_VM look_VVI towards_II the_AT nearest_JJT door_NN1 available_JJ for_IF being_VBG shut_VVN ._. 
(_( If_CS that_DD1 door_NN1 is_VBZ shut_VVN ,_, he_PPHS1 may_VM well_RR say_VVI '_" It_PPH1 's_VBZ shut_JJ '_GE ,_, rather_CS21 than_CS22 consider_VVI what_DDQ other_JJ doors_NN2 are_VBR potentially_RR available_JJ for_IF being_VBG shut_VVN ._. )_) 
Similarly_RR if_CS his_APPGE host_NN1 says_58 '_GE Come_VV0 early_RR '_GE ,_, having_VHG just_RR invited_VVN him_PPHO1 for_IF eight_MC o'clock_RA ,_, he_PPHS1 will_VM interpret_VVI '_GE early_JJ '_GE with_II31 respect_II32 to_II33 the_AT last-mentioned_JJ time_NNT1 ,_, rather_CS21 than_CS22 to_II some_DD previously_RR mentioned_VVN time_NNT1 ._. 
Consider_VV0 again_RT extract_VV0 (_( 17_MC )_) presented_VVD here_RL as_CSA (_( 27_MC )_) ._. 
(_( 27_MC )_) a_AT1 man_NN1 and_CC woman_NN1 sitting_VVG in_II the_AT living_NN1 room_NN1 ..._... the_AT man_NN1 's_GE bored_JJ goes_VVZ to_II the_AT window_NN1 looks_VVZ out_RP the_AT window_NN1 ..._... and_CC goes_VVZ out_RP +_FO goes_VVZ to_II his_PPGE goes_VVZ to_II a_AT1 club_NN1 +_FO has_VHZ a_AT1 drink_NN1 talks_NN2 to_II the_AT barman_NN1 In_II our_APPGE discussion_NN1 in_II 2.2.2_MC ,_, we_PPIS2 pointed_VVD out_RP the_AT effect_NN1 of_IO '_GE co-text_NN1 '_GE in_II limiting_VVG the_AT interpretation_NN1 of_IO what_DDQ follows_VVZ ._. 
The_AT initial_JJ setting_NN1 of_IO the_AT co-text_NN1 determines_VVZ the_AT extent_NN1 of_IO the_AT context_NN1 within_II which_DDQ the_AT hearer_NN1 will_VM understand_VVI what_DDQ is_VBZ said_VVN next_MD ._. 
He_PPHS1 assumes_VVZ that_CST entities_NN2 referred_VVN to_II will_VM remain_VVI constant_JJ ,_, that_CST the_AT temporal_JJ setting_NN1 will_VM remain_VVI constant_JJ ,_, that_CST the_AT locational_JJ setting_NN1 will_VM remain_VVI constant_JJ ,_, unless_CS the_AT speaker_NN1 indicates_VVZ some_DD change_NN1 in_II any_DD of_IO these_DD2 ,_, in_II which_DDQ case_VV0 the_AT hearer_NN1 will_VM minimally_RR expand_VVI the_AT context_NN1 ._. 
Not_XX only_RR does_VDZ the_AT hearer_NN1 assume_VVI it_PPH1 is_VBZ the_AT same_DA '_GE man_NN1 '_GE who_PNQS is_VBZ being_VBG talked_VVN about_II throughout_RL ,_, he_PPHS1 also_RR assumes_VVZ that_CST the_AT man_NN1 will_VM stay_VVI in_II the_AT same_DA place_NN1 unless_CS the_AT speaker_NN1 announces_VVZ that_CST he_PPHS1 moves_VVZ ._. 
When_CS the_AT hearer_NN1 hears_VVZ goes_VVZ to_II the_AT window_NN1 ,_, he_PPHS1 assumes_VVZ it_PPH1 is_58 '_GE the_AT window_NN1 '_GE in_II that_DD1 same_DA '_GE living_NN1 room_NN1 '_GE which_DDQ has_VHZ already_RR been_VBN mentioned_VVN ,_, and_CC he_PPHS1 assumes_VVZ that_CST the_AT man_NN1 '_GE goes_VVZ to_II the_AT window_NN1 '_GE on_II the_AT same_DA occasion_NN1 ,_, within_II minutes_NNT2 of_IO the_AT original_JJ setting_NN1 '_GE sitting_NN1 in_II the_AT living_NN1 room_NN1 '_GE ._. 
When_CS the_AT man_NN1 goes_VVZ to_II a_AT1 club_NN1 ,_, the_AT hearer_NN1 assumes_VVZ that_CST the_AT '_GE club_NN1 '_GE is_VBZ in_II the_AT same_DA town_NN1 ,_, that_CST the_AT man_NN1 has_VHZ not_XX caught_VVN an_AT1 aeroplane_NN1 and_CC flown_VVN to_II Las_NP1 Vegas_NP1 ._. 
Again_RT the_AT minimal_JJ expansion_NN1 of_IO the_AT spatio-temporal_JJ setting_NN1 will_VM suggest_VVI that_CST the_AT man_NN1 has_VHZ a_AT1 drink_NN1 and_CC talks_NN2 to_II the_AT barman_NN1 within_II that_DD1 same_DA club_NN1 and_CC on_II that_DD1 same_DA occasion_NN1 ,_, within_II a_AT1 restricted_JJ time-span_NN1 ,_, say_VV0 an_AT1 hour_NNT1 rather_II21 than_II22 a_AT1 year_NNT1 ._. 
It_PPH1 is_VBZ this_DD1 principle_NN1 ,_, which_DDQ instructs_VVZ the_AT hearer_NN1 not_XX to_TO construct_VVI a_AT1 context_NN1 any_DD larger_JJR than_CSN necessary_JJ to_TO secure_VVI an_AT1 interpretation_NN1 ,_, which_DDQ accounts_VVZ for_IF how_RRQ we_PPIS2 understand_VV0 Sacks_NP1 '_GE (_( 1972_MC )_) much-quoted_JJ sequence_NN1 :_: (_( 28_MC )_) The_AT baby_NN1 cried_VVD ._. 
The_AT mommy_NN1 picked_VVD it_PPH1 up_RP ._. 
It_PPH1 is_VBZ possible_JJ ,_, of_RR21 course_RR22 ,_, to_TO imagine_VVI that_CST the_AT first_MD of_IO these_DD2 sentences_NN2 describes_VVZ one_MC1 event_NN1 and_CC the_AT second_MD describes_VVZ another_DD1 ,_, quite_RG unrelated_JJ ,_, event_NN1 (_( so_RR the_AT person_NN1 identified_VVD as_58 '_GE a_AT1 mother_NN1 '_GE may_VM be_VBI picking_VVG up_RP a_AT1 chair_NN1 in_II the_AT course_NN1 of_IO cleaning_VVG a_AT1 room_NN1 )_) ._. 
The_AT principle_NN1 of_IO local_JJ interpretation_NN1 however_RR ,_, will_VM guide_VVI us_PPIO2 to_TO construct_VVI a_AT1 limited_JJ context_NN1 in_II which_DDQ '_VBZ the_AT mother_NN1 '_GE is_VBZ the_AT mentioned_VVN baby_NN1 's_GE mother_NN1 and_CC the_AT expression_NN1 it_PPH1 is_VBZ used_VVN to_TO refer_VVI to_II the_AT previously_RR mentioned_VVN baby_NN1 ._. 
Moreover_RR the_AT sequence_NN1 of_IO events_NN2 will_VM be_VBI understood_VVN as_CSA happening_VVG adjacently_RR in_II time_NNT1 and_CC situated_VVN adjacently_RR in_II place_NN1 ._. 
It_PPH1 does_VDZ not_XX even_RR occur_VVI to_II the_AT reader_NN1 that_CST the_AT baby_NN1 might_VM have_VHI cried_VVN one_MC1 year_NNT1 in_II Singapore_NP1 and_CC be_VBI picked_VVN up_RP by_II its_APPGE mother_NN1 a_AT1 year_NNT1 later_RRR in_II Aden_NP1 ._. 
It_PPH1 would_VM ,_, of_RR21 course_RR22 ,_, be_VBI possible_JJ to_TO establish_VVI a_AT1 setting_NN1 in_II which_DDQ such_DA a_AT1 sequence_NN1 of_IO events_NN2 would_VM be_VBI plausible_JJ ,_, but_CCB ,_, if_CS no_AT such_DA setting_NN1 is_VBZ established_VVN ,_, the_AT reader_NN1 will_VM assume_VVI a_AT1 local_JJ interpretation_NN1 in_II31 respect_II32 of_II33 time_NNT1 ,_, place_NN1 and_CC participants_NN2 ._. 
It_PPH1 must_VM be_VBI obvious_JJ that_DD1 '_VBZ local_JJ interpretation_NN1 '_GE may_VM only_RR be_VBI vaguely_RR conceptualised_VVN ._. 
It_PPH1 seems_VVZ unlikely_JJ that_CST in_II interpreting_VVG (_( 28_MC )_) the_AT reader_NN1 postulates_VVZ any_DD exact_JJ physical_JJ distance_NN1 between_II the_AT mother_NN1 and_CC the_AT baby_NN1 at_II the_AT point_NN1 before_II the_AT mother_NN1 picks_VVZ the_AT child_NN1 up_RP ,_, or_CC that_CST he_PPHS1 bothers_VVZ to_TO wonder_VVI whether_CSW the_AT mother_NN1 picks_VVZ the_AT child_NN1 up_RP after_CS it_PPH1 has_VHZ finished_VVN crying_NN1 (_( and_CC if_CS so_RR how_RGQ long_RR after_CS ,_, in_II31 terms_II32 of_II33 minutes_NNT2 or_CC seconds_NNT2 )_) or_CC whether_CSW the_AT child_NN1 was_VBDZ still_RR crying_VVG when_RRQ the_AT mother_NN1 picked_VVD it_PPH1 up_RP ._. 
Similarly_RR it_PPH1 seems_VVZ unlikely_JJ that_CST the_AT reader_NN1 will_VM bother_VVI to_TO construct_VVI a_AT1 three-dimensional_JJ ,_, photographic_JJ representation_NN1 of_IO '_GE the_AT baby_NN1 '_GE which_DDQ cries_VVZ in_II the_AT first_MD sentence_NN1 and_CC which_DDQ is_VBZ picked_VVN up_RP in_II the_AT second_MD sentence_NN1 ._. 
'_GE Local_JJ interpretation_NN1 '_GE probably_RR relates_VVZ to_II another_DD1 strategy_NN1 which_DDQ instructs_VVZ the_AT hearer_NN1 /_FO reader_NN1 to_TO do_VDI as_RG little_DA1 processing_NN1 as_CSA possible_JJ ,_, only_RR to_TO construct_VVI a_AT1 representation_NN1 which_DDQ is_VBZ sufficiently_RR specific_JJ to_TO permit_VVI an_AT1 interpretation_NN1 which_DDQ is_VBZ adequate_JJ for_IF what_DDQ the_AT hearer_NN1 judges_VVZ the_AT purpose_NN1 of_IO the_AT utterance_NN1 to_TO be_VBI ._. 
Everything_PN1 that_CST we_PPIS2 have_VH0 said_VVN so_RG far_RR in_II this_DD1 section_NN1 leans_VVZ heavily_RR on_II the_AT hearer_NN1 's_VBZ /_FO reader_NN1 's_GE ability_NN1 to_TO utilise_VVI his_APPGE knowledge_NN1 of_IO the_AT world_NN1 and_CC his_APPGE past_JJ experience_NN1 of_IO similar_JJ events_NN2 in_II interpreting_VVG the_AT language_NN1 which_DDQ he_PPHS1 encounters_VVZ ._. 
It_PPH1 is_VBZ the_AT experience_NN1 of_IO similar_JJ events_NN2 which_DDQ enables_VVZ him_PPHO1 to_TO judge_VVI what_DDQ the_AT purpose_NN1 of_IO an_AT1 utterance_NN1 might_VM be_VBI ._. 
It_PPH1 is_VBZ his_APPGE knowledge_NN1 of_IO the_AT world_NN1 which_DDQ constrains_VVZ his_APPGE local_JJ interpretation_NN1 ._. 
Consider_VV0 again_RT (_( 27_MC )_) presented_VVD here_RL as_CSA (_( 29_MC )_) ._. 
(_( 29_MC )_) a_AT1 man_NN1 and_CC woman_NN1 sitting_VVG in_II the_AT living_NN1 room_NN1 ..._... the_AT man_NN1 's_GE bored_JJ goes_VVZ to_II the_AT window_NN1 ..._... goes_VVZ out_RP ..._... goes_VVZ to_II a_AT1 club_NN1 We_PPIS2 suggested_VVD that_CST goes_VVZ to_II the_AT window_NN1 will_VM be_VBI interpreted_VVN as_CSA meaning_VVG that_DD1 '_VBZ he_PPHS1 goes_VVZ to_II the_AT window_NN1 in_II the_AT living_NN1 room_NN1 '_GE ,_, whereas_CS goes_NN2 to_II a_AT1 club_NN1 will_VM be_VBI interpreted_VVN as_CSA meaning_NN1 '_GE goes_VVZ to_II a_AT1 club_NN1 in_II the_AT same_DA town_NN1 '_GE ,_, i.e._REX not_XX '_GE in_II the_AT living_NN1 room_NN1 '_GE ,_, nor_CC even_RR '_GE in_II the_AT same_DA house_NN1 '_GE ._. 
Knowledge_NN1 of_IO the_AT world_NN1 tells_VVZ us_PPIO2 that_CST houses_NN2 which_DDQ contain_VV0 living_JJ rooms_NN2 do_VD0 not_XX usually_RR contain_VVI bars_NN2 ._. 
Goes_VVZ out_RP can_VM not_XX be_VBI simply_RR interpreted_VVN as_CSA meaning_NN1 '_GE goes_VVZ out_II21 of_II22 the_AT room_NN1 '_GE ,_, it_PPH1 has_VHZ to_TO be_VBI interpreted_VVN as_CSA meaning_NN1 '_GE goes_VVZ out_II21 of_II22 the_AT house_NN1 '_GE ._. 
(_( In_II Chapter_NN1 7_MC we_PPIS2 return_VV0 to_II a_AT1 discussion_NN1 of_IO '_GE knowledge_NN1 of_IO the_AT world_NN1 '_GE ._. )_) 
We_PPIS2 must_VM suppose_VVI that_CST an_AT1 individual_NN1 's_GE experience_NN1 of_IO past_JJ events_NN2 of_IO a_AT1 similar_JJ kind_NN1 will_VM equip_VVI him_PPHO1 with_IW expectations_NN2 ,_, hypotheses_NN2 ,_, about_II what_DDQ are_VBR likely_JJ to_TO be_VBI relevant_JJ aspects_NN2 of_IO context_NN1 ._. 
Bartlett_NP1 ,_, one_MC1 of_IO the_AT founders_NN2 of_IO modern_JJ psychology_NN1 ,_, comments_NN2 on_II the_AT importance_NN1 of_IO relating_VVG a_AT1 particular_JJ experience_NN1 to_II other_JJ similar_JJ experiences_NN2 :_: it_PPH1 is_VBZ legitimate_JJ to_TO say_VVI that_CST all_DB the_AT cognitive_JJ processes_NN2 which_DDQ have_VH0 been_VBN considered_VVN ,_, from_II perceiving_VVG to_II thinking_VVG ,_, are_VBR ways_NN2 in_II which_DDQ some_DD fundamental_JJ '_GE effort_NN1 after_II meaning_NN1 '_GE seeks_VVZ expression_NN1 ._. 
Speaking_VVG very_RG broadly_RR ,_, such_DA effort_NN1 is_VBZ simply_RR the_AT attempt_NN1 to_TO connect_VVI something_PN1 that_CST is_VBZ given_VVN with_IW something_PN1 other_II21 than_II22 itself._NNU (_( 1932_MC :_: 227_MC ,_, our_APPGE emphasis_NN1 )_) The_AT individual_NN1 ,_, he_PPHS1 suggests_VVZ ,_, generalises_VVZ over_RP particular_JJ experiences_NN2 and_CC extracts_NN2 from_II these_DD2 a_AT1 number_NN1 of_IO types_NN2 of_IO experience_NN1 ._. 
This_DD1 notion_NN1 is_VBZ ,_, of_RR21 course_RR22 ,_, implicit_JJ in_II the_AT construction_NN1 of_IO the_AT sets_NN2 of_IO features_NN2 of_IO context_NN1 which_DDQ we_PPIS2 have_VH0 been_VBN considering_VVG in_II this_DD1 chapter_NN1 ._. 
In_BCL21 order_BCL22 to_TO construct_VVI a_AT1 notion_NN1 of_IO '_GE speaker_NN1 in_II a_AT1 context_NN1 '_GE it_PPH1 is_VBZ necessary_JJ to_TO generalise_VVI over_RP contexts_NN2 and_CC to_TO determine_VVI what_DDQ characteristics_NN2 speakers_NN2 in_II different_JJ contexts_NN2 share_VV0 ._. 
Similarly_RR ,_, in_BCL21 order_BCL22 to_TO construct_VVI a_AT1 notion_NN1 of_IO '_GE genre_NN1 '_GE ,_, it_PPH1 is_VBZ necessary_JJ to_TO generalise_VVI across_II experience_NN1 and_CC determine_VVI what_DDQ it_PPH1 is_VBZ that_DD1 is_VBZ common_JJ to_II fairy_NN1 stories_NN2 ,_, chats_NN2 ,_, news_NN1 broadcasts_NN2 ,_, epic_JJ poems_NN2 ,_, debates_NN2 or_CC salesmen_NN2 's_GE routines_NN2 which_DDQ enables_VVZ us_PPIO2 to_TO recognise_VVI one_PN1 as_CSA being_VBG a_AT1 token_NN1 of_IO the_AT generalised_JJ type_NN1 ._. 
On_II the_AT basis_NN1 of_IO experience_NN1 then_RT ,_, we_PPIS2 recognise_VV0 types_NN2 of_IO communicative_JJ events_NN2 which_DDQ take_VV0 place_NN1 against_II the_AT background_NN1 of_IO a_AT1 mass_NN1 of_IO below-conscious_JJ expectations_NN2 also_RR based_VVN on_II past_JJ experience_NN1 which_DDQ we_PPIS2 might_VM summarise_VVI ,_, following_VVG van_NP1 Dijk_NP1 (_( 1977_MC :_: 99_MC )_) ,_, as_58 '_GE the_AT ASSUMED_JJ NORMALITY_NN1 of_IO the_AT world_NN1 ._. 
We_PPIS2 assume_VV0 that_CST our_APPGE muscles_NN2 will_VM continue_VVI to_TO move_VVI normally_RR ,_, that_CST doors_NN2 which_DDQ normally_RR open_VV0 will_VM continue_VVI to_TO open_VVI ,_, that_DD1 hair_NN1 grows_VVZ on_II heads_NN2 ,_, that_CST dogs_NN2 bark_VV0 ,_, that_CST towns_NN2 retain_VV0 their_APPGE geographical_JJ locations_NN2 ,_, that_CST the_AT sun_NN1 will_VM shine_VVI ,_, and_RR31 so_RR32 on_RR33 ._. 
It_PPH1 is_VBZ interesting_JJ to_TO observe_VVI the_AT powerful_JJ constraints_NN2 on_II creators_NN2 of_IO surrealist_JJ or_CC science_NN1 fiction_NN1 in_II this_DD1 respect_NN1 ._. 
Alice_NP1 may_VM enter_VVI a_AT1 looking-glass_NN1 world_NN1 where_CS unexpected_JJ things_NN2 happen_VV0 ,_, but_CCB she_PPHS1 is_VBZ still_RR constituted_VVN like_II a_AT1 human_JJ being_NN1 :_: walking_NN1 may_VM take_VVI her_PPHO1 in_II an_AT1 unexpected_JJ direction_NN1 ,_, but_CCB the_AT nature_NN1 of_IO the_AT physical_JJ act_NN1 of_IO walking_NN1 is_VBZ taken_VVN for_IF granted_VVN ._. 
If_CS too_RG many_DA2 expectations_NN2 are_VBR flouted_VVN ,_, the_AT writer_NN1 may_VM be_VBI suspected_VVN of_IO being_VBG mentally_RR unbalanced_JJ ,_, of_IO being_VBG incapable_JJ of_IO seeing_VVG the_AT world_NN1 in_II a_AT1 normal_JJ way_NN1 ._. 
Thus_RR ,_, on_II the_AT one_MC1 hand_NN1 ,_, expectations_NN2 make_VV0 interpretation_NN1 possible_JJ and_CC ,_, on_II the_AT other_JJ ,_, they_PPHS2 constitute_VV0 an_AT1 extension_NN1 or_CC further_JJR affirmation_NN1 of_IO their_APPGE own_DA validity_NN1 ._. 
Popper_NN1 makes_VVZ the_AT point_NN1 cogently_RR :_: '_" we_PPIS2 are_VBR born_VVN with_IW expectations_NN2 :_: with_IW '_GE knowledge_NN1 '_GE which_DDQ ,_, although_CS not_XX valid_JJ a_JJ21 priori_JJ22 ,_, is_VBZ psychologically_RR orgenetically_RR a_JJ21 priori_JJ22 ,_, i.e._REX prior_II21 to_II22 all_DB observational_JJ experience_NN1 ._. 
One_MC1 of_IO the_AT most_RGT important_JJ of_IO these_DD2 expectations_NN2 is_VBZ the_AT expectation_NN1 of_IO finding_VVG a_AT1 regularity_NN1 ._. 
It_PPH1 is_VBZ connected_VVN with_IW an_AT1 inborn_JJ propensity_NN1 to_TO look_VVI out_RP for_IF regularities_NN2 ,_, or_CC with_IW a_AT1 need_NN1 to_TO find_VVI regularities_NN2 '_GE (_( 1963_MC :_: 47_MC ,_, original_JJ emphasis_NN1 )_) ._. 
Furthermore_RR ,_, as_CSA Lewis_NP1 (_( 1969_MC :_: 38_MC )_) points_VVZ out_RP ,_, '_GE fortunately_RR we_PPIS2 have_VH0 learned_VVN that_CST all_DB of_IO us_PPIO2 will_VM mostly_RR notice_VVI the_AT same_DA analogies_NN2 '_GE ._. 
Not_XX only_RR are_VBR we_PPIS2 all_DB primed_VVN to_TO look_VVI for_IF regularities_NN2 ,_, we_PPIS2 tend_VV0 to_TO perceive_VVI the_AT same_DA regularities_NN2 ._. 
Clearly_RR the_AT smaller_JJR the_AT community_NN1 ,_, the_AT more_DAR notions_NN2 of_IO regularity_NN1 will_VM be_VBI shared_VVN ,_, since_CS the_AT contexts_NN2 which_DDQ the_AT members_NN2 of_IO the_AT community_NN1 share_NN1 will_VM be_VBI very_RG similar_JJ ._. 
Once_CS the_AT individual_NN1 begins_VVZ to_TO establish_VVI regularities_NN2 ,_, to_TO generalise_VVI over_II experience_NN1 ,_, it_PPH1 becomes_VVZ possible_JJ for_IF him_PPHO1 not_XX only_RR to_TO recognise_VVI a_AT1 particular_JJ experience_NN1 as_CSA being_VBG one_MC1 of_IO a_AT1 type_NN1 ,_, say_VV0 a_AT1 scolding_NN1 or_CC an_AT1 interview_NN1 ,_, it_PPH1 also_RR becomes_VVZ possible_JJ to_TO predict_VVI what_DDQ is_VBZ likely_JJ to_TO happen_VVI ,_, what_DDQ are_VBR likely_JJ to_TO be_VBI the_AT relevant_JJ features_NN2 of_IO context_NN1 ,_, within_II a_AT1 particular_JJ type_NN1 of_IO communicative_JJ event_NN1 ._. 
It_PPH1 follows_VVZ that_CST the_AT hearer_NN1 in_II a_AT1 speech_NN1 situation_NN1 is_VBZ not_XX in_II the_AT position_NN1 of_IO trying_VVG to_TO pay_VVI attention_NN1 to_II every_AT1 feature_NN1 of_IO the_AT context_NN1 (_( in_II principle_NN1 an_AT1 impossible_JJ task_NN1 )_) ._. 
He_PPHS1 only_RR pays_VVZ attention_NN1 to_II those_DD2 features_NN2 which_DDQ have_VH0 been_VBN necessary_JJ and_CC relevant_JJ in_II similar_JJ situations_NN2 in_II the_AT past_NN1 ._. 
Bartlett_NP1 suggests_VVZ that_CST the_AT individual_JJ has_58 '_GE an_AT1 overmastering_JJ tendency_NN1 simply_RR to_TO get_VVI a_AT1 general_JJ impression_NN1 of_IO the_AT whole_NN1 ;_; and_CC on_II the_AT basis_NN1 of_IO this_DD1 he_PPHS1 constructs_NN2 the_AT probable_JJ detail_NN1 '_GE (_( 1932_MC :_: 206_MC )_) ._. 
We_PPIS2 pay_VV0 attention_NN1 to_II those_DD2 salient_JJ features_NN2 which_DDQ are_VBR constitutive_JJ of_IO the_AT type_NN1 of_IO genre_NN1 ,_, and_CC expect_VV0 that_CST the_AT peripheral_JJ features_NN2 will_VM be_VBI as_CSA they_PPHS2 have_VH0 been_VBN in_II the_AT past_NN1 ._. 
Obviously_RR there_EX will_VM be_VBI types_NN2 of_IO occasions_NN2 which_DDQ have_VH0 not_XX occurred_VVN within_II our_APPGE past_JJ experience_NN1 ._. 
We_PPIS2 have_VH0 cultural_JJ stereotypes_NN2 which_DDQ suggest_VV0 that_CST such_DA occasions_NN2 are_VBR difficult_JJ for_IF us_PPIO2 ,_, potentially_RR embarrassing_JJ ,_, because_CS we_PPIS2 do_VD0 not_XX know_VVI the_AT appropriate_JJ responses_NN2 ._. 
Thus_RR ,_, if_CS it_PPH1 is_VBZ the_AT first_MD time_NNT1 someone_PN1 tells_VVZ you_PPY a_AT1 particular_JJ genre_NN1 of_IO joke_NN1 ,_, you_PPY may_VM not_XX know_VVI the_AT appropriate_JJ type_NN1 of_IO response_NN1 ._. 
The_AT second_MD time_NNT1 around_RP ,_, however_RR ,_, you_PPY feel_VV0 more_RGR confident_JJ of_IO what_DDQ to_TO expect_VVI ._. 
(_( Tolstoy_NP1 ,_, in_II War_NN1 and_CC Peace_NN1 ,_, gives_VVZ a_AT1 brilliant_JJ account_NN1 of_IO the_AT insecurity_NN1 engendered_VVN by_II the_AT first_MD occasion_NN1 of_IO a_AT1 new_JJ type_NN1 of_IO experience_NN1 in_II his_APPGE description_NN1 of_IO Pierre_NP1 's_GE induction_NN1 into_II membership_NN1 of_IO a_AT1 masonic_JJ brotherhood_NN1 ._. )_) 
Our_APPGE experience_NN1 of_IO particular_JJ communicative_JJ situations_NN2 teaches_VVZ us_PPIO2 what_DDQ to_TO expect_VVI of_IO that_DD1 situation_NN1 ,_, both_RR in_II a_AT1 general_JJ predictive_JJ sense_NN1 (_( e.g._REX the_AT sort_NN1 of_IO attitudes_NN2 which_DDQ are_VBR likely_JJ to_TO be_VBI expressed_VVN ,_, the_AT sort_NN1 of_IO topics_NN2 which_DDQ are_VBR likely_JJ to_TO be_VBI raised_VVN )_) which_DDQ gives_VVZ rise_NN1 to_II notions_NN2 of_IO '_GE appropriacy_NN1 '_GE ,_, and_CC in_II a_AT1 limited_JJ predictive_JJ sense_NN1 which_DDQ enables_VVZ us_PPIO2 to_TO interpret_VVI linguistic_JJ tokens_NN2 (_( e.g._REX deictic_JJ forms_NN2 like_II here_RL and_CC now_RT )_) in_II the_AT way_NN1 we_PPIS2 have_VH0 interpreted_VVN them_PPHO2 before_RT in_II similar_JJ contexts_NN2 ._. 
We_PPIS2 must_VM assume_VVI that_CST the_AT young_JJ child_NN1 's_GE acquisition_NN1 of_IO language_NN1 comes_VVZ about_RP in_II the_AT context_NN1 of_IO expanding_JJ experience_NN1 ,_, of_IO expanding_VVG possible_JJ interpretations_NN2 of_IO forms_NN2 like_II here_RL and_CC now_RT in_II different_JJ contexts_NN2 of_IO situation_NN1 ,_, contexts_NN2 which_DDQ come_VV0 to_TO be_VBI recognised_VVN ,_, and_CC stored_JJ as_CSA types_NN2 ._. 
Against_II the_AT background_NN1 of_IO this_DD1 mass_NN1 of_IO expectations_NN2 which_DDQ derives_VVZ from_II and_CC constitutes_VVZ our_APPGE experience_NN1 ,_, it_PPH1 must_VM become_VVI possible_JJ to_TO identify_VVI the_AT relevant_JJ properties_NN2 of_IO features_NN2 of_IO the_AT context_NN1 of_IO situation_NN1 in_II31 terms_II32 of_II33 norms_NN2 of_IO expectation_NN1 within_II a_AT1 particular_JJ genre_NN1 ._. 
The_AT more_RGR highly_RR constrained_VVN and_CC ritualised_VVN the_AT genre_NN1 ,_, the_AT more_RGR likely_JJ we_PPIS2 are_VBR to_TO be_VBI able_JK to_TO identify_VVI norms_NN2 ._. 
Thus_RR it_PPH1 seems_VVZ likely_JJ that_CST examination_NN1 questions_NN2 in_II chemical_JJ engineering_NN1 at_II degree_NN1 level_NN1 will_VM bear_VVI certain_JJ similarities_NN2 of_IO form_NN1 and_CC content_NN1 ,_, and_CC share_VV0 certain_JJ presuppositions_NN2 ,_, in_II institutions_NN2 throughout_II the_AT world_NN1 ._. 
The_AT less_RRR constrained_VVD the_AT genre_NN1 ,_, primarily_RR interactional_JJ '_GE chat_NN1 '_GE ,_, for_REX21 example_REX22 ,_, the_AT less_RGR likely_JJ it_PPH1 is_VBZ that_CST we_PPIS2 can_VM confidently_RR state_VVI norms_NN2 of_IO expectation_NN1 which_DDQ will_VM generalise_VVI even_RR over_II the_AT experience_NN1 of_IO the_AT English-speaking_JJ population_NN1 ._. 
For_IF the_AT individual_JJ participant_NN1 in_II a_AT1 chatting_NN1 relationship_NN1 ,_, this_DD1 does_VDZ not_XX constitute_VVI a_AT1 difficulty_NN1 ,_, because_CS he_PPHS1 has_VHZ plenty_PN of_IO previous_JJ personal_JJ and_CC local_JJ experience_NN1 to_TO call_VVI upon_II ._. 
For_IF the_AT discourse_NN1 analyst_NN1 ,_, on_II the_AT other_JJ hand_NN1 ,_, the_AT more_RGR personal_JJ and_CC particular_JJ the_AT occasion_NN1 for_IF the_AT participants_NN2 ,_, the_AT more_RGR limited_JJ and_CC circumspect_JJ he_PPHS1 must_VM be_VBI in_II his_APPGE interpretation_NN1 ._. 
Confronted_VVN with_IW data_NN of_IO the_AT following_JJ sort_NN1 ,_, an_AT1 extract_NN1 from_II a_AT1 private_JJ diary_NN1 only_RR intended_VVN to_TO remind_VVI the_AT elderly_JJ writer_NN1 of_IO how_RRQ she_PPHS1 passed_VVD a_AT1 day_NNT1 in_II January_NPM1 1982_MC ,_, the_AT discourse_NN1 analyst_NN1 may_VM not_XX be_VBI able_JK to_TO proceed_VVI very_RG far_RR in_II his_APPGE analysis._NNU (_( 30_MC )_) Did_VDD more_DAR to_II Ivy_NP1 's_GE letter_NN1 ._. 
A.A._NP1 rang_VVD me_PPIO1 at_II 4_MC o/c_ZZ1 she_PPHS1 returned_VVD on_II 2nd_MD and_CC had_VHD had_VHN grand_JJ time_NNT1 with_IW Gwenda_NP1 and_CC families_NN2 ._. 
As_CSA was_VBDZ nice_RR p.m_RA ._. 
I_PPIS1 went_VVD to_II Evensong_NP1 (_( rev_NNU ._. 
Carlil_VV0 )_) and_CC walked_VVD back_RP with_IW Mrs._NNB Nicholls_NP1 (_( 85_MC !_! !_! )_) and_CC daughter_NN1 ._. 
Cos_NP2@ '_GE Doris_NP1 rang_VVD 8.15_MC and_CC will_VM come_VVI tomorrow_RT !_! 
Bed._NP1 11.15_MC ._. 
Of_RR21 course_RR22 ,_, if_CS the_AT discourse_NN1 analyst_NN1 experiences_NN2 a_AT1 great_JJ deal_NN1 of_IO data_NN like_II this_DD1 ,_, he_PPHS1 will_VM feel_VVI more_RGR confident_JJ in_II his_APPGE description_NN1 and_CC interpretation_NN1 ._. 
He_PPHS1 ,_, too_RR ,_, is_VBZ constrained_VVN in_II his_APPGE interpretation_NN1 by_II past_JJ similar_JJ experience_NN1 ,_, by_II interpreting_VVG in_II41 the_II42 light_II43 of_II44 what_DDQ we_PPIS2 might_VM call_VVI the_AT principle_NN1 of_IO analogy_NN1 ._. 
The_AT principle_NN1 of_IO analogy_NN1 will_VM provide_VVI a_AT1 reasonably_RR secure_JJ framework_NN1 for_IF interpretation_NN1 for_IF the_AT hearer_NN1 and_CC for_IF the_AT analyst_NN1 most_DAT of_IO the_AT time_NNT1 ._. 
Most_DAT of_IO the_AT time_NNT1 ,_, things_NN2 will_VM indeed_RR conform_VVI to_II our_APPGE expectations_NN2 ._. 
However_RR ,_, conventions_NN2 can_VM be_VBI flouted_VVN and_CC expectations_NN2 upset_VV0 ,_, either_RR deliberately_RR for_IF a_AT1 stylistic_JJ effect_NN1 ,_, or_CC by_II accident_NN1 or_CC oversight_NN1 ._. 
Note_VV0 that_CST where_CS the_AT speaker_NN1 /_FO writer_NN1 is_VBZ deliberately_RR flouting_VVG a_AT1 convention_NN1 ,_, upsetting_VVG an_AT1 expectation_NN1 for_IF a_AT1 stylistic_JJ effect_NN1 ,_, he_PPHS1 can_VM only_RR bring_VVI off_RP that_DD1 effect_NN1 because_CS the_AT convention_NN1 /_FO expectation_NN1 exists_VVZ ._. 
The_AT '_GE non-limerick_NN1 '_GE which_DDQ follows_VVZ only_RR makes_VVZ an_AT1 effect_NN1 in_II41 the_II42 light_II43 of_II44 the_AT conventional_JJ structure_NN1 for_IF limericks_NN2 which_DDQ have_VH0 a_AT1 characteristic_JJ rhythm_NN1 and_CC an_AT1 aabba_NN1 rhyme_NN1 scheme_NN1 :_: (_( 31_MC )_) There_EX was_VBDZ a_AT1 young_JJ girl_NN1 of_IO St_NP1 Bees_NP1 ,_, Who_PNQS was_VBDZ stung_VVN on_II the_AT nose_NN1 by_II a_AT1 wasp_NN1 When_RRQ asked_VVN '_" Does_VDZ it_PPH1 hurt_VVI ?_? 
'_" She_PPHS1 replied_VVD '_" Yes_UH it_PPH1 does_VDZ ,_, But_CCB I_PPIS1 'm_VBM glad_JJ it_PPH1 was_VBDZ n't_XX a_AT1 hornet_NN1 ._. 
'_GE The_AT principle_NN1 of_IO analogy_NN1 is_VBZ one_MC1 of_IO the_AT fundamental_JJ heuristics_NN2 which_DDQ hearers_NN2 and_CC analysts_NN2 adopt_VV0 in_II determining_JJ interpretations_NN2 in_II context_NN1 ._. 
They_PPHS2 assume_VV0 that_CST everything_PN1 will_VM remain_VVI as_CSA it_PPH1 was_VBDZ before_CS unless_CS they_PPHS2 are_VBR given_VVN specific_JJ notice_NN1 that_CST some_DD aspect_NN1 has_VHZ changed_VVN ._. 
Dahl_NP1 (_( 1976_MC :_: 46_MC )_) formulates_VVZ a_AT1 principle_NN1 for_IF speakers_NN2 :_: '_GE Indicate_VV0 only_JJ things_NN2 which_DDQ have_VH0 changed_VVN and_CC omit_VV0 those_DD2 which_DDQ are_VBR as_CSA they_PPHS2 were_VBDR before_RT ._. 
'_" To_TO repeat_VVI what_DDQ is_VBZ known_VVN to_TO be_VBI shared_JJ knowledge_NN1 ,_, '_GE things_NN2 as_CSA they_PPHS2 were_VBDR before_RT '_GE ,_, flouts_VVZ Grice_NP1 's_GE maxim_NN1 of_IO quantity_NN1 ._. 
(_( Speakers_NN2 do_VD0 ,_, of_RR21 course_RR22 ,_, remind_VVI each_PPX221 other_PPX222 of_IO knowledge_NN1 which_DDQ they_PPHS2 share_VV0 ,_, in_BCL21 order_BCL22 to_TO make_VVI that_DD1 knowledge_NN1 part_NN1 of_IO the_AT activated_JJ context_NN1 of_IO discourse_NN1 ,_, as_CSA McCawley_NP1 (_( 1979_MC )_) points_VVZ out_RP ._. )_) 
Discourse_NN1 is_VBZ interpreted_VVN in_II41 the_II42 light_II43 of_II44 past_JJ experience_NN1 of_IO similar_JJ discourse_NN1 ,_, by_II analogy_NN1 with_IW previous_JJ similar_JJ texts_NN2 (_( remember_VV0 the_AT relevance_NN1 of_IO experience_NN1 of_IO previous_JJ similar_JJ texts_NN2 in_II the_AT interpretation_NN1 of_IO (_( 14C_FO )_) in_II Chapter_NN1 2_MC ,_, SQUASHED_JJ INSECTS_NN2 DO_VD0 NT_NP1 BITE_VVI MAD_JJ MENTAL_JJ RULE_NN1 )_) ._. 
Relevant_JJ previous_JJ experience_NN1 ,_, together_RL with_IW the_AT principle_NN1 of_IO local_JJ interpretation_NN1 ,_, will_VM impel_VVI hearers_NN2 /_FO readers_NN2 to_TO try_VVI to_TO interpret_VVI sequential_JJ utterances_NN2 as_CSA relating_VVG to_II the_AT same_DA topic_NN1 ._. 
When_CS two_MC sentences_NN2 are_VBR placed_VVN together_RL in_II sequence_NN1 by_II a_AT1 writer_NN1 who_PNQS does_VDZ not_XX want_VVI us_PPIO2 to_TO consider_VVI them_PPHO2 as_II a_AT1 continuous_JJ text_NN1 ,_, their_APPGE separateness_NN1 or_CC disconnectedness_NN1 must_VM be_VBI positively_RR indicated_VVN ._. 
In_II a_AT1 linguistics_NN1 textbook_NN1 ,_, the_AT following_JJ two_MC sentences_NN2 were_VBDR presented_VVN as_CSA separate_JJ citation_NN1 examples_NN2 to_TO illustrate_VVI structural_JJ ambiguity._NNU (_( 32_MC )_) 1_MC1 ._. 
The_AT bride_NN1 and_CC groom_VV0 left_JJ early_JJ last_MD night._NNT1 2_MC ._. 
He_PPHS1 greeted_VVD the_AT girl_NN1 with_IW a_AT1 smile_NN1 ._. 
(_( Brown_NP1 &amp;_CC Miller_NP1 ,_, 1980_MC :_: 84_MC )_) In_II the_AT context_NN1 of_IO a_AT1 linguistics_NN1 textbook_NN1 ,_, especially_RR one_MC1 on_II syntax_NN1 ,_, we_PPIS2 would_VM not_XX expect_VVI to_TO have_VHI to_TO interpret_VVI two_MC continuous_JJ cited_JJ sentences_NN2 as_CSA describing_VVG an_AT1 event_NN1 sequence_NN1 ._. 
In_II most_DAT contexts_NN2 ,_, however_RR ,_, the_AT natural_JJ '_GE effort_NN1 after_II meaning_NN1 '_GE will_NN1 impel_VV0 the_AT hearer_NN1 /_FO reader_NN1 to_TO try_VVI to_TO cointerpret_VVI chunks_NN2 of_IO language_NN1 which_DDQ he_PPHS1 finds_VVZ close_RR to_II each_PPX221 other_PPX222 on_II a_AT1 page_NN1 ,_, or_CC a_AT1 stone_NN1 or_CC a_AT1 wall_NN1 and_CC ,_, where_RRQ possible_JJ ,_, to_TO interpret_VVI the_AT language_NN1 as_CSA relevant_JJ to_II the_AT physical_JJ context_NN1 ._. 
This_DD1 last_MD point_NN1 leads_VVZ us_PPIO2 to_II an_AT1 important_JJ ,_, but_CCB frequently_RR misunderstood_VVN ,_, concept_NN1 in_II the_AT analysis_NN1 of_IO discourse_NN1 ._. 
The_AT imperative_JJ 'need_NN1 to_TO find_VVI regularities_NN2 '_GE which_DDQ Popper_NN1 speaks_VVZ of_IO ,_, coupled_VVN with_IW Bartlett_NP1 's_GE '_GE effort_NN1 after_II meaning_NN1 '_GE ,_, constitute_VV0 a_AT1 powerful_JJ expectation_NN1 in_II human_JJ beings_NN2 that_CST what_DDQ is_VBZ said_VVN or_CC written_VVN will_VM make_VVI sense_NN1 in_II the_AT context_NN1 in_II which_DDQ it_PPH1 appears_VVZ ._. 
Even_RR in_II the_AT most_RGT unpropitious_JJ circumstances_NN2 ,_, the_AT natural_JJ reaction_NN1 of_IO man_NN1 appears_VVZ to_TO be_VBI to_TO make_VVI sense_NN1 of_IO any_DD sign_NN1 resembling_VVG language_NN1 ,_, resembling_VVG an_AT1 effort_NN1 to_TO communicate_VVI ._. 
The_AT reaction_NN1 of_IO the_AT man_NN1 who_PNQS finds_VVZ what_DDQ are_VBR apparently_RR signs_NN2 etched_VVN in_II a_AT1 stone_NN1 in_II the_AT middle_NN1 of_IO a_AT1 desert_NN1 is_VBZ to_TO try_VVI to_TO decipher_VVI their_APPGE meaning_NN1 ._. 
The_AT reaction_NN1 of_IO parents_NN2 to_II infants_NN2 ,_, and_CC of_IO friends_NN2 to_II the_AT speech_NN1 of_IO those_DD2 who_PNQS are_VBR gravely_RR ill_JJ ,_, is_VBZ to_TO attribute_VVI meaning_VVG to_II any_DD murmur_NN1 which_DDQ can_VM be_VBI interpreted_VVN as_CSA relevant_JJ to_II the_AT context_NN1 of_IO situation_NN1 ,_, and_CC ,_, if_CS at_RR21 all_RR22 possible_JJ ,_, to_TO interpret_VVI what_DDQ appears_VVZ to_TO be_VBI being_VBG said_VVN as_CSA constituting_VVG a_AT1 coherent_JJ message_NN1 ,_, permitting_VVG the_AT hearer_NN1 to_TO construct_VVI a_AT1 coherent_JJ interpretation_NN1 ._. 
The_AT natural_JJ effort_NN1 of_IO hearers_NN2 and_CC readers_NN2 alike_RR is_VBZ to_TO attribute_VVI relevance_NN1 and_CC coherence_NN1 to_II the_AT text_NN1 they_PPHS2 encounter_VV0 until_CS they_PPHS2 are_VBR forced_VVN not_XX to_TO ._. 
The_AT normal_JJ expectation_NN1 in_II the_AT construction_NN1 and_CC interpretation_NN1 of_IO discourse_NN1 is_VBZ ,_, as_CSA Grice_NP1 suggests_VVZ ,_, that_DD1 relevance_NN1 holds_VVZ ,_, that_CST the_AT speaker_NN1 is_VBZ still_RR speaking_VVG of_IO the_AT same_DA place_NN1 and_CC time_NNT1 ,_, participants_NN2 and_CC topic_NN1 ,_, unless_CS he_PPHS1 marks_VVZ a_AT1 change_NN1 and_CC shows_VVZ explicitly_RR whether_CSW the_AT changed_JJ context_NN1 is_VBZ ,_, or_CC is_VBZ not_XX ,_, relevant_JJ to_II what_DDQ he_PPHS1 has_VHZ been_VBN saying_VVG previously_RR ._. 
Similarly_RR the_AT normal_JJ expectation_NN1 is_VBZ that_CST the_AT discourse_NN1 will_VM be_VBI coherent_JJ ._. 
The_AT reaction_NN1 of_IO some_DD scholars_NN2 to_II the_AT question_NN1 of_IO '_GE coherence_NN1 '_GE is_VBZ to_TO search_VVI for_IF cues_NN2 to_II coherence_NN1 within_II the_AT text_NN1 and_CC this_DD1 may_VM indeed_RR yield_VVI a_AT1 descriptive_JJ account_NN1 of_IO the_AT characteristics_NN2 of_IO some_DD types_NN2 of_IO text_NN1 ._. 
It_PPH1 ignores_VVZ ,_, however_RR ,_, the_AT fact_NN1 that_CST human_JJ beings_NN2 do_VD0 not_XX require_VVI formal_JJ textual_JJ markers_NN2 before_CS they_PPHS2 are_VBR prepared_VVN to_TO interpret_VVI a_AT1 text_NN1 ._. 
They_PPHS2 naturally_RR assume_VV0 coherence_NN1 ,_, and_CC interpret_VV0 the_AT text_NN1 in_II41 the_II42 light_II43 of_II44 that_DD1 assumption_NN1 ._. 
They_PPHS2 assume_VV0 ,_, that_REX21 is_REX22 ,_, that_CST the_AT principles_NN2 of_IO analogy_NN1 and_CC local_JJ interpretation_NN1 constrain_VV0 their_APPGE experience_NN1 ._. 
There_EX are_VBR as_RG many_DA2 linguistic_JJ '_GE cues_NN2 to_II coherence_NN1 '_GE (_( a_AT1 concept_NN1 to_TO be_VBI discussed_VVN in_II detail_NN1 in_II Chapter_NN1 6_MC )_) holding_VVG between_II the_AT pairs_NN2 of_IO sentences_NN2 :_: (_( 33_MC )_) 1_MC1 ._. 
The_AT bride_NN1 and_CC groom_VV0 left_JJ early_JJ last_MD night._NNT1 2_MC ._. 
He_PPHS1 greeted_VVD the_AT girl_NN1 with_IW a_AT1 smile._NNU as_CSA there_EX are_VBR between_II :_: (_( 34_MC )_) The_AT baby_NN1 cried_VVD ._. 
The_AT mommy_NN1 picked_VVD it_PPH1 up_RP ._. 
It_PPH1 is_VBZ not_XX the_AT sequence_NN1 of_IO sentences_NN2 which_DDQ represents_58 '_GE coherent_JJ discourse_NN1 '_GE ._. 
Rather_RR it_PPH1 is_VBZ the_AT reader_NN1 ,_, driven_VVN by_II the_AT principles_NN2 of_IO analogy_NN1 and_CC local_JJ interpretation_NN1 ,_, who_PNQS assumes_VVZ that_CST the_AT second_MD sequence_NN1 describes_VVZ a_AT1 series_NN of_IO connected_JJ events_NN2 and_CC interprets_VVZ linguistic_JJ cues_NN2 (_( like_II baby_NN1 it_PPH1 )_) under_II that_DD1 assumption_NN1 ._. 
Encountering_VVG the_AT first_MD pair_NN of_IO sentences_NN2 in_II the_AT context_NN1 in_II which_DDQ they_PPHS2 occur_VV0 ,_, the_AT reader_NN1 does_VDZ not_XX assume_VVI that_CST they_PPHS2 describe_VV0 a_AT1 connected_JJ sequence_NN1 of_IO events_NN2 and_CC consequently_RR does_VDZ not_XX interpret_VVI the_AT potential_JJ linguistic_JJ cues_NN2 (_( like_II groom_NN1 he_PPHS1 )_) as_CSA referring_VVG to_II the_AT same_DA entity_NN1 ._. 
The_AT principles_NN2 of_IO analogy_NN1 (_( things_NN2 will_VM tend_VVI to_TO be_VBI as_CSA they_PPHS2 were_VBDR before_II )_) and_CC local_JJ interpretation_NN1 (_( if_CS there_EX is_VBZ a_AT1 change_NN1 ,_, assume_VV0 it_PPH1 is_VBZ minimal_JJ )_) form_VV0 the_AT basis_NN1 of_IO the_AT assumption_NN1 of_IO coherence_NN1 in_II our_APPGE experience_NN1 of_IO life_NN1 in_RR21 general_RR22 ,_, hence_RR in_II our_APPGE experience_NN1 of_IO discourse_NN1 as_RR21 well._RR22 3_MC Topic_NN1 and_CC the_AT representation_NN1 of_IO discourse_NN1 content_NN1 In_II the_AT course_NN1 of_IO this_DD1 chapter_NN1 ,_, we_PPIS2 shall_VM examine_VVI some_DD of_IO the_AT uses_NN2 of_IO the_AT term_NN1 topic_NN1 in_II the_AT study_NN1 of_IO discourse_NN1 ._. 
In_II the_AT process_NN1 ,_, we_PPIS2 shall_VM explore_VVI some_DD recent_JJ attempts_NN2 to_TO construct_VVI a_AT1 theoretical_JJ notion_NN1 of_IO '_GE topic_NN1 '_GE ,_, a_AT1 notion_NN1 which_DDQ seems_VVZ to_TO be_VBI essential_JJ to_II concepts_NN2 such_II21 as_II22 '_GE relevance_NN1 '_GE and_CC '_GE coherence_NN1 '_GE ,_, but_CCB which_DDQ itself_PPX1 is_VBZ very_RG difficult_JJ to_TO pin_VVI down_RP ._. 
We_PPIS2 shall_VM suggest_VVI that_DD1 formal_JJ attempts_NN2 to_TO identify_VVI topics_NN2 are_VBR doomed_VVN to_II failure_NN1 ,_, but_CCB that_CST the_AT discourse_NN1 analyst_NN1 may_VM usefully_RR make_VVI appeal_NN1 to_II notions_NN2 like_NN1 '_GE speaking_VVG topically_RR '_GE and_CC '_GE the_AT speaker_NN1 's_GE topic_NN1 '_GE within_II a_AT1 '_GE topic_NN1 framework_NN1 '_GE ._. 
We_PPIS2 shall_VM also_RR consider_VVI briefly_RR how_RRQ markers_NN2 of_IO '_GE topic-shift_NN1 '_GE may_VM be_VBI identified_VVN in_II written_JJ and_CC spoken_JJ discourse_NN1 ._. 
In_RR21 particular_RR22 ,_, we_PPIS2 shall_VM insist_VVI on_II the_AT principle_NN1 that_CST it_PPH1 is_VBZ speakers_NN2 and_CC writers_NN2 who_PNQS have_VH0 topics_NN2 ,_, not_XX texts_NN2 ._. 
We_PPIS2 shall_VM then_RT go_VVI on_RP to_TO consider_VVI how_RRQ the_AT notion_NN1 of_IO '_GE topic_NN1 '_GE relates_VVZ to_II representations_NN2 of_IO discourse_NN1 content_NN1 ._. 
Since_CS many_DA2 of_IO the_AT representations_NN2 proposed_VVN are_VBR based_VVN on_II a_AT1 hierarchical_JJ organisation_NN1 of_IO discourse_NN1 content_NN1 ,_, we_PPIS2 shall_VM consider_VVI critically_RR the_AT possibility_NN1 of_IO characterising_NN1 '_GE topic_NN1 '_GE in_II31 terms_II32 of_II33 the_AT top-most_JJ elements_NN2 in_II the_AT hierarchical_JJ representation_NN1 ._. 
Discourse_NN1 fragments_NN2 and_CC the_AT notion_NN1 '_GE topic'_NN1 We_PPIS2 have_VH0 already_RR argued_VVN that_CST the_AT data_NN used_VVN in_II discourse_NN1 analysis_NN1 will_VM inevitably_RR reflect_VVI the_AT analyst_NN1 's_GE particular_JJ interests_NN2 ._. 
Moreover_RR ,_, the_AT piece_NN1 of_IO data_NN chosen_VVN for_IF study_NN1 can_VM only_RR be_VBI partially_RR analysed_VVN ._. 
If_CS the_AT investigation_NN1 is_VBZ undertaken_VVN by_II someone_PN1 primarily_RR interested_JJ in_II intonation_NN1 ,_, for_REX21 example_REX22 ,_, the_AT data_NN selected_JJ has_VHZ to_TO meet_VVI certain_JJ requirements_NN2 ._. 
It_PPH1 must_VM be_VBI spoken_VVN ,_, audible_JJ ,_, and_CC ,_, depending_II21 on_II22 the_AT level_NN1 of_IO investigation_NN1 involved_VVD ,_, clear_RR enough_RR to_TO allow_VVI instrumental_JJ analysis_NN1 ,_, and_CC accompanied_VVN by_II additional_JJ information_NN1 on_II the_AT age_NN1 ,_, sex_NN1 and_CC linguistic_JJ background_NN1 of_IO the_AT speaker_NN1 ._. 
In_II practice_NN1 ,_, any_DD single_JJ investigation_NN1 will_VM have_VHI much_DA1 stricter_JJR data_NN requirements_NN2 than_CSN this_DD1 rather_RG general_JJ list_NN1 ._. 
Having_VHG selected_VVN the_AT data_NN ,_, the_AT investigators_NN2 will_VM study_VVI features_NN2 such_II21 as_II22 the_AT pitch_NN1 ,_, rhythm_NN1 and_CC loudness_NN1 of_IO syllables_NN2 in_II the_AT data_NN ,_, and_CC spend_VV0 relatively_RR little_JJ or_CC no_AT time_NNT1 studying_VVG the_AT lexis_NN1 or_CC the_AT morphology_NN1 ._. 
In_II its_APPGE most_RGT extreme_JJ form_NN1 ,_, this_DD1 narrowing_NN1 of_IO the_AT investigation_NN1 in_II31 terms_II32 of_II33 the_AT data_NN selected_VVN and_CC the_AT analysis_NN1 undertaken_VVN can_VM lead_VVI to_II a_AT1 constructed_JJ text_NN1 being_VBG carefully_RR read_VVN aloud_RR in_II a_AT1 phonetics_NN1 laboratory_NN1 by_II a_AT1 speaker_NN1 of_IO standard_JJ Southern_JJ British_JJ English_NN1 ._. 
The_AT results_NN2 of_IO the_AT investigation_NN1 may_VM then_RT be_VBI used_VVN to_TO make_VVI '_GE empirical_JJ '_GE claims_NN2 about_II the_AT intonation_NN1 of_IO English_NN1 ._. 
Although_CS this_DD1 is_VBZ an_AT1 extreme_JJ example_NN1 ,_, it_PPH1 serves_VVZ to_TO illustrate_VVI the_AT selectiveness_NN1 which_DDQ characterises_VVZ linguistic_JJ investigation_NN1 generally_RR ,_, and_CC which_DDQ is_VBZ also_RR present_JJ to_II a_AT1 certain_JJ degree_NN1 in_II most_DAT analysis_NN1 of_IO discourse_NN1 ._. 
The_AT data_NN studied_VVN in_II discourse_NN1 analysis_NN1 is_VBZ always_RR a_AT1 fragment_NN1 of_IO discourse_NN1 and_CC the_AT discourse_NN1 analyst_NN1 always_RR has_VHZ to_TO decide_VVI where_RRQ the_AT fragment_NN1 begins_VVZ and_CC ends_NN2 ._. 
How_RRQ does_VDZ the_AT analyst_NN1 decide_VVI what_DDQ constitutes_VVZ a_AT1 satisfactory_JJ unit_NN1 for_IF analysis_NN1 ?_? 
There_RL do_VD0 exist_VVI ways_NN2 of_IO identifying_VVG the_AT boundaries_NN2 of_IO stretches_NN2 of_IO discourse_NN1 which_DDQ set_VVD one_MC1 chunk_NN1 of_IO discourse_NN1 off_RP from_II the_AT rest_NN1 ._. 
Formulaic_JJ expressions_NN2 such_II21 as_II22 '_GE Once_RR upon_II a_AT1 time_NNT1 ..._... and_CC they_PPHS2 lived_VVD happily_RR ever_RR after_CS '_" can_VM be_VBI used_VVN explicitly_RR to_TO mark_VVI the_AT boundaries_NN2 of_IO a_AT1 fragment_NN1 ._. 
Other_JJ familiar_JJ markers_NN2 are_VBR '_" Have_VH0 you_PPY heard_VVN the_AT one_PN1 about_II ..._... ?_? 
'_GE ,_, '_" Did_VDD I_PPIS1 tell_VVI you_PPY what_DDQ happened_VVD to_II me_PPIO1 last_MD week._NNT1 ?_? '_GE and_CC various_JJ other_JJ forms_NN2 which_DDQ can_VM be_VBI used_VVN to_TO mark_VVI the_AT beginning_NN1 of_IO a_AT1 joke_NN1 or_CC anecdote_NN1 ._. 
These_DD2 markers_NN2 can_VM help_VVI the_AT analyst_NN1 decide_VV0 where_RRQ the_AT beginning_NN1 of_IO a_AT1 coherent_JJ fragment_NN1 of_IO discourse_NN1 occurs_VVZ ._. 
However_RR ,_, speakers_NN2 often_RR do_VD0 not_XX provide_VVI such_DA explicit_JJ guidelines_NN2 to_TO help_VVI the_AT analyst_NN1 select_JJ chunks_NN2 of_IO discourse_NN1 for_IF study_NN1 ._. 
In_BCL21 order_BCL22 to_TO divide_VVI up_RP a_AT1 lengthy_JJ recording_NN1 of_IO conversational_JJ data_NN into_II chunks_NN2 which_DDQ can_VM be_VBI investigated_VVN in_II detail_NN1 ,_, the_AT analyst_NN1 is_VBZ often_RR forced_VVN to_TO depend_VVI on_II intuitive_JJ notions_NN2 about_II where_RRQ one_MC1 part_NN1 of_IO a_AT1 conversation_NN1 ends_NN2 and_CC another_DD1 begins_VVZ ._. 
There_EX are_VBR ,_, of_RR21 course_RR22 ,_, points_VVZ where_RRQ one_MC1 speaker_NN1 stops_VVZ and_CC another_DD1 starts_VVZ speaking_VVG ,_, but_CCB every_AT1 speaker-change_NN1 does_VDZ not_XX necessarily_RR terminate_VVI a_AT1 particular_JJ coherent_JJ fragment_NN1 of_IO conversation_NN1 ._. 
Which_DDQ point_NN1 of_IO speaker-change_NN1 ,_, among_II the_AT many_DA2 ,_, could_VM be_VBI treated_VVN as_II the_AT end_NN1 of_IO one_MC1 chunk_NN1 of_IO the_AT conversation_NN1 ?_? 
This_DD1 type_NN1 of_IO decision_NN1 is_VBZ typically_RR made_VVN by_II appealing_VVG to_II an_AT1 intuitive_JJ notion_NN1 of_IO topic_NN1 ._. 
The_AT conversationalists_NN2 stop_VV0 talking_VVG about_II '_GE money_NN1 '_GE and_CC move_VV0 on_RP to_II '_GE sex_NN1 '_GE ._. 
A_AT1 chunk_NN1 of_IO conversational_JJ discourse_NN1 ,_, then_RT ,_, can_VM be_VBI treated_VVN as_II a_AT1 unit_NN1 of_IO some_DD kind_NN1 because_CS it_PPH1 is_VBZ on_II a_AT1 particular_NN1 '_GE topic_NN1 '_GE ._. 
The_AT notion_NN1 of_IO '_GE topic_NN1 '_GE is_VBZ clearly_RR an_AT1 intuitively_RR satisfactory_JJ way_NN1 of_IO describing_VVG the_AT unifying_JJ principle_NN1 which_DDQ makes_VVZ one_MC1 stretch_NN1 of_IO discourse_NN1 '_GE about_II '_" something_PN1 and_CC the_AT next_MD stretch_NN1 '_GE about_II '_" something_PN1 else_RR ,_, for_IF it_PPH1 is_VBZ appealed_VVN to_II very_RG frequently_RR in_II the_AT discourse_NN1 analysis_NN1 literature_NN1 ._. 
Yet_RR the_AT basis_NN1 for_IF the_AT identification_NN1 of_IO '_GE topic_NN1 '_GE is_VBZ rarely_RR made_VVN explicit_JJ ._. 
In_II fact_NN1 ,_, '_GE topic_NN1 '_GE could_VM be_VBI described_VVN as_II the_AT most_RGT frequently_RR used_VVN ,_, unexplained_JJ ,_, term_NN1 in_II the_AT analysis_NN1 of_IO discourse_NN1 ._. 
Sentential_JJ topic_NN1 One_MC1 use_NN1 of_IO the_AT term_NN1 '_GE topic_NN1 '_GE is_VBZ associated_VVN with_IW descriptions_NN2 of_IO sentence_NN1 structure_NN1 ._. 
According_II21 to_II22 Hockett_NP1 ,_, a_AT1 distinction_NN1 can_VM be_VBI made_VVN between_II the_AT topic_NN1 and_CC the_AT comment_NN1 in_II a_AT1 sentence_NN1 ,_, in_II that_DD1 '_VBZ the_AT speaker_NN1 announces_VVZ a_AT1 topic_NN1 and_CC then_RT says_VVZ something_PN1 about_II it_PPH1 ..._... 
In_II English_NN1 and_CC the_AT familiar_JJ languages_NN2 of_IO Europe_NP1 ,_, topics_NN2 are_VBR usually_RR also_RR subjects_NN2 and_CC comments_NN2 are_VBR predicates_NN2 '_GE (_( 1958_MC :_: 201_MC )_) ._. 
It_PPH1 is_VBZ clear_JJ from_II Hockett_NP1 's_GE examples_NN2 ,_, reproduced_VVD here_RL as_CSA (_( 1_MC1 )_) and_CC (_( 2_MC )_) ,_, that_DD1 this_58 '_GE sentential_JJ topic_NN1 '_GE may_VM coincide_VVI with_IW the_AT grammatical_JJ subject_NN1 ,_, as_CSA in_II (_( 1_MC1 )_) ,_, but_CCB need_VM not_XX ,_, as_CSA in_II (_( 2_MC )_) ._. 
(_( 1_MC1 )_) John_NP1 /_FO ran_VVD away_RL (_( 2_MC )_) That_CST new_JJ book_NN1 by_II Thomas_NP1 Guernsey_NP1 /_FO I_ZZ1 have_VH0 n't_XX read_VVN yet_RR The_AT treatment_NN1 of_IO '_GE topic_NN1 '_GE as_II a_AT1 grammatical_JJ term_NN1 ,_, identifying_VVG a_AT1 constituent_NN1 in_II the_AT structure_NN1 of_IO a_AT1 sentence_NN1 (_( or_CC the_AT deep_JJ structure_NN1 analysis_NN1 ,_, at_RR21 least_RR22 )_) is_VBZ also_RR noticeable_JJ in_II the_AT work_NN1 of_IO grammarians_NN2 such_II21 as_II22 Dahl_NP1 (_( 1969_MC )_) and_CC Sgall_NP1 et_RA21 al._RA22 (_( 1973_MC )_) ._. 
Transformational_JJ generative_JJ grammars_NN2 would_VM also_RR account_VVI for_IF the_AT structure_NN1 of_IO example_NN1 (_( 2_MC )_) in_II31 terms_II32 of_II33 a_AT1 movement_NN1 transformation_NN1 called_VVN '_GE topicalisation_NN1 '_GE ._. 
The_AT term_NN1 '_GE topic_NN1 '_GE ,_, then_RT ,_, as_CSA found_VVN in_II descriptions_NN2 of_IO sentence_NN1 structure_NN1 ,_, is_VBZ essentially_RR a_AT1 term_NN1 which_DDQ identifies_VVZ a_AT1 particular_JJ sentential_JJ constituent_NN1 ._. 
As_II such_DA ,_, it_PPH1 has_VHZ been_VBN used_VVN in_II the_AT study_NN1 of_IO discourse_NN1 ,_, by_II Grimes_NP1 (_( 1975_MC :_: 337_MC )_) for_REX21 example_REX22 ,_, to_TO describe_VVI the_AT different_JJ methods_NN2 used_VVN in_II various_JJ languages_NN2 to_TO mark_VVI the_AT '_GE topic_NN1 constituent_NN1 '_GE of_IO sentences_NN2 ._. 
It_PPH1 has_VHZ also_RR been_VBN used_VVN by_II Givn_NP1 (_( 1979a_FO )_) in_II his_APPGE argument_NN1 that_CST ,_, in_II the_AT development_NN1 of_IO a_AT1 language_NN1 ,_, sentential_JJ subjects_NN2 are_VBR derived_VVN from_II '_GE grammaticalised_JJ topics_NN2 '_GE ._. 
However_RR ,_, we_PPIS2 are_VBR not_XX ,_, for_IF the_AT moment_NN1 ,_, concerned_JJ with_IW the_AT structure_NN1 of_IO linguistic_JJ units_NN2 comparable_JJ to_II the_AT simple_JJ sentence_NN1 (_( see_VV0 Chapter_NN1 5_MC )_) ._. 
Nor_CC are_VBR we_PPIS2 considering_RR '_GE topic_NN1 '_GE as_II a_AT1 grammatical_JJ constituent_NN1 of_IO any_DD kind_NN1 ._. 
We_PPIS2 are_VBR primarily_RR interested_JJ in_II the_AT general_JJ pretheoretical_JJ notion_NN1 of_IO '_GE topic_NN1 '_GE as_58 '_GE what_DDQ is_VBZ being_VBG talked_VVN about_II '_GE in_II a_AT1 conversation_NN1 ._. 
This_DD1 type_NN1 of_IO '_GE topic_NN1 '_GE is_VBZ unlikely_JJ to_TO be_VBI identifiable_JJ as_CSA one_MC1 part_NN1 of_IO a_AT1 sentence_NN1 ._. 
Accordingly_RR ,_, we_PPIS2 agree_VV0 with_IW Morgan_NP1 that_DD1 '_VBZ it_PPH1 is_VBZ not_XX sentences_NN2 that_CST have_VH0 topics_NN2 ,_, but_CCB speakers_NN2 '_GE (_( Morgan_NP1 ,_, 1975_MC :_: 434_MC )_) ._. 
Discourse_NN1 topic_NN1 In_II an_AT1 attempt_NN1 to_TO distinguish_VVI their_APPGE notion_NN1 of_IO topic_NN1 from_II the_AT grammarians_NN2 '_GE sentential_JJ topic_NN1 ,_, Keenan_NP1 &amp;_CC Schieffelin_NP1 (_( 1976_MC )_) used_VVD the_AT term_NN1 discourse_NN1 topic_NN1 ._. 
They_PPHS2 were_VBDR particularly_RR anxious_JJ to_TO avoid_VVI having_VHG '_GE topic_NN1 '_GE ,_, in_II discourse_NN1 study_NN1 ,_, treated_VVN as_CS21 if_CS22 it_PPH1 were_VBDR somehow_RR expressible_JJ by_II a_AT1 simple_JJ noun_NN1 phrase_NN1 ,_, as_CSA often_RR happens_VVZ in_II the_AT treatment_NN1 of_IO sentential_JJ topics_NN2 ._. 
(_( Some_DD ontological_JJ reasons_NN2 for_IF this_DD1 type_NN1 of_IO treatment_NN1 are_VBR suggested_VVN by_II Lyons_NP1 ,_, 1977_MC :_: 502_MC ._. )_) 
What_DDQ Keenan_NP1 &amp;_CC Schieffelin_NP1 (_( 1976_MC :_: 380_MC )_) emphasise_VV0 is_VBZ that_DD1 '_VBZ discourse_NN1 topic_NN1 is_VBZ not_XX a_AT1 simple_JJ NP_NP1 ,_, but_CCB a_AT1 proposition_NN1 (_( about_II which_DDQ some_DD claim_NN1 is_VBZ made_VVN or_CC elicited_VVN )_) '_GE ._. 
It_PPH1 may_VM be_VBI because_CS their_APPGE investigation_NN1 is_VBZ primarily_RR concerned_JJ with_IW children_NN2 's_GE speech_NN1 ,_, but_CCB ,_, in_II describing_VVG the_AT discourse_NN1 topic_NN1 as_II the_AT '_GE question_NN1 of_IO immediate_JJ concern_NN1 '_GE ,_, Keenan_NP1 &amp;_CC Schieffelin_NP1 appear_VV0 to_TO replace_VVI the_AT idea_NN1 of_IO a_AT1 single_JJ correct_JJ noun_NN1 phrase_NN1 as_CSA expressing_VVG the_AT topic_NN1 with_IW the_AT idea_NN1 of_IO a_AT1 single_JJ correct_JJ phrase_NN1 or_CC sentence_NN1 ._. 
The_AT implication_NN1 in_II their_APPGE study_NN1 is_VBZ that_CST there_EX must_VM be_VBI ,_, for_IF any_DD fragment_NN1 of_IO conversational_JJ discourse_NN1 ,_, a_AT1 single_JJ proposition_NN1 (_( expressed_VVN as_II a_AT1 phrase_NN1 or_CC sentence_NN1 )_) which_DDQ represents_VVZ the_AT discourse_NN1 topic_NN1 of_IO the_AT whole_NN1 of_IO the_AT fragment_NN1 ._. 
Such_DA a_AT1 view_NN1 is_VBZ certainly_RR too_RG simplistic_JJ ,_, as_CSA we_PPIS2 hope_VV0 to_TO show_VVI by_II considering_VVG some_DD experimental_JJ work_NN1 in_II which_DDQ '_VBZ the_AT topic_NN1 '_GE was_VBDZ treated_VVN as_II the_AT equivalent_NN1 of_IO a_AT1 title_NN1 ._. 
(_( We_PPIS2 shall_VM consider_VVI the_AT possibility_NN1 of_IO representing_JJ '_GE the_AT discourse_NN1 topic_NN1 '_GE as_II a_AT1 proposition_NN1 when_RRQ we_PPIS2 investigate_VV0 the_AT proposition-based_JJ analysis_NN1 of_IO discourse_NN1 in_II section_NN1 3.7_MC ._. )_) 
In_II a_AT1 series_NN of_IO experiments_NN2 reported_VVN by_II Bransford_NP1 &amp;_CC Johnson_NP1 (_( 1973_MC )_) subjects_NN2 were_VBDR presented_VVN with_IW constructed_JJ texts_NN2 to_TO read_VVI ,_, comprehend_VV0 ,_, and_CC ,_, later_RRR ,_, recall_VV0 ._. 
The_AT aim_NN1 of_IO the_AT experiments_NN2 was_VBDZ to_TO demonstrate_VVI that_CST the_AT comprehension_NN1 of_IO English_JJ texts_NN2 depends_VVZ not_XX only_RR on_II knowledge_NN1 of_IO the_AT language_NN1 ,_, but_CCB also_RR on_II extra-linguistic_JJ knowledge_NN1 ,_, particularly_RR related_VVN to_II the_AT contexts_NN2 in_II which_DDQ the_AT texts_NN2 occur_VV0 ._. 
There_EX are_VBR examples_NN2 of_IO texts_NN2 which_DDQ appear_VV0 to_TO depend_VVI on_II accompanying_VVG visual_JJ material_NN1 for_IF comprehension_NN1 and_CC others_NN2 ,_, such_II21 as_II22 example_NN1 (_( 3_MC )_) reproduced_VVD below_RL ,_, for_IF which_DDQ '_VBZ the_AT topic_NN1 '_GE must_VM be_VBI provided._NNU (_( 3_MC )_) The_AT procedure_NN1 is_VBZ actually_RR quite_RG simple_JJ ._. 
First_MD you_PPY arrange_VV0 things_NN2 into_II different_JJ groups_NN2 ._. 
Of_RR21 course_RR22 ,_, one_MC1 pile_NN1 may_VM be_VBI sufficient_JJ depending_II21 on_II22 how_RGQ much_RR there_EX is_VBZ to_TO do_VDI ._. 
If_CS you_PPY have_VH0 to_TO go_VVI somewhere_RL else_RR due_JJ to_TO lack_VVI of_IO facilities_NN2 that_CST is_VBZ the_AT next_MD step_NN1 ,_, otherwise_RR you_PPY are_VBR pretty_RG well_RR set_VVN ._. 
It_PPH1 is_VBZ important_JJ not_XX to_TO overdo_VVI things_NN2 ._. 
That_REX21 is_REX22 ,_, it_PPH1 is_VBZ better_JJR to_TO do_VDI too_RR few_DA2 things_NN2 at_RR21 once_RR22 than_CSN too_RG many_DA2 ._. 
In_II the_AT short_JJ run_NN1 this_DD1 may_VM not_XX seem_VVI important_JJ but_CCB complications_NN2 can_VM easily_RR arise_VVI ._. 
A_AT1 mistake_NN1 can_VM be_VBI expensive_JJ as_RR21 well_RR22 ._. 
At_RR21 first_RR22 the_AT whole_JJ procedure_NN1 will_VM seem_VVI complicated_JJ ._. 
Soon_RR ,_, however_RR ,_, it_PPH1 will_VM become_VVI just_RR another_DD1 facet_NN1 of_IO life_NN1 ._. 
It_PPH1 is_VBZ difficult_JJ to_TO foresee_VVI any_DD end_NN1 to_II the_AT necessity_NN1 for_IF this_DD1 task_NN1 in_II the_AT immediate_JJ future_NN1 ,_, but_CCB then_RT one_PN1 never_RR can_VM tell_VVI ._. 
After_CS the_AT procedure_NN1 is_VBZ completed_VVN one_PN1 arranges_VVZ the_AT materials_NN2 into_II different_JJ groups_NN2 again_RT ._. 
Then_RT they_PPHS2 can_VM be_VBI put_VVN into_II their_APPGE appropriate_JJ places_NN2 ._. 
Eventually_RR they_PPHS2 will_VM be_VBI used_VVN once_RR21 more_RR22 and_CC the_AT whole_JJ cycle_NN1 will_VM then_RT have_VHI to_TO be_VBI repeated_VVN ._. 
However_RR ,_, that_DD1 is_VBZ part_NN1 of_IO (_( from_II Bransford_NP1 &amp;_CC Johnson_NP1 ,_, 1973_MC :_: 400_MC )_) Because_CS it_PPH1 was_VBDZ constructed_VVN for_IF a_AT1 specific_JJ purpose_NN1 ,_, this_DD1 text_NN1 is_VBZ fairly_RR unusual_JJ in_CS21 that_CS22 there_EX are_VBR few_DA2 lexical_JJ clues_NN2 to_II what_DDQ the_AT text_NN1 might_VM be_VBI '_GE about_II '_GE ._. 
Predictably_RR ,_, the_AT experiments_NN2 showed_VVD that_DD1 comprehension_NN1 and_CC recall_NN1 of_IO this_DD1 passage_NN1 were_VBDR significantly_RR better_RRR when_CS subjects_NN2 were_VBDR provided_VVN ,_, before_II reading_NN1 ,_, with_IW what_DDQ Bransford_NP1 &amp;_CC Johnson_NP1 called_JJ '_GE the_AT topic_NN1 of_IO the_AT passage_NN1 '_GE ._. 
The_AT topic_NN1 of_IO this_DD1 passage_NN1 was_58 '_GE Washing_NN1 clothes_NN2 '_GE ._. 
The_AT reader_NN1 can_VM judge_VVI for_IF himself_PPX1 whether_CSW his_APPGE comprehension_NN1 would_VM have_VHI been_VBN fuller_JJR if_CS he_PPHS1 had_VHD known_VVN this_DD1 topic_NN1 ._. 
The_AT use_NN1 of_IO the_AT word_NN1 '_GE topic_NN1 '_GE in_II this_DD1 type_NN1 of_IO experiment_NN1 suggests_VVZ that_CST the_AT topic_NN1 of_IO a_AT1 text_NN1 is_VBZ equivalent_JJ to_II the_AT title_NN1 and_CC that_DD1 ,_, for_IF any_DD text_NN1 ,_, there_EX is_VBZ a_AT1 single_JJ correct_JJ expression_NN1 which_DDQ is_58 '_GE the_AT topic_NN1 '_GE ._. 
This_DD1 would_VM be_VBI the_AT case_NN1 if_CS texts_NN2 could_VM only_RR be_VBI understood_VVN completely_RR as_CS31 long_CS32 as_CS33 they_PPHS2 were_VBDR accompanied_VVN by_II the_AT single_JJ ,_, correct_JJ title_NN1 ._. 
However_RR ,_, it_PPH1 should_VM not_XX be_VBI too_RG difficult_JJ to_TO imagine_VVI several_DA2 different_JJ titles_NN2 for_IF passage_NN1 (_( 3_MC )_) ,_, each_DD1 of_IO which_DDQ could_VM equally_RR facilitate_VVI comprehension_NN1 ._. 
One_PN1 could_VM indicate_VVI that_CST the_AT text_NN1 contains_VVZ a_AT1 set_NN1 of_IO instructions_NN2 by_II producing_VVG a_AT1 title_NN1 such_II21 as_II22 '_GE How_RRQ to_TO Do_VDI the_AT Laundry_NN1 '_GE or_CC '_" A_ZZ1 Guide_VV0 to_II Getting_VVG your_APPGE Clothes_NN2 Cleaner_NN1 '_GE ._. 
One_PN1 could_VM incorporate_VVI the_AT text_NN1 's_GE philosophical_JJ final_JJ statement_NN1 in_II a_AT1 title_NN1 such_II21 as_II22 '_GE Doing_VDG the_AT Laundry_NN1 as_II a_AT1 Philosophy_NN1 of_IO Life_NN1 '_GE or_CC '_GE An_AT1 Orderly_JJ Life_NN1 through_II Good_JJ Laundry_NN1 Procedure_NN1 '_GE ._. 
These_DD2 latter_DA titles_NN2 contain_VV0 as_RG much_DA1 information_NN1 for_IF the_AT reader_NN1 as_CSA the_AT title_NN1 '_GE Washing_NN1 Clothes_NN2 '_GE ,_, which_DDQ Bransford_NP1 &amp;_CC Johnson_NP1 describe_VV0 as_II '_GE the_AT topic_NN1 '_GE ._. 
The_AT implication_NN1 ,_, surely_RR ,_, is_VBZ that_DD1 ,_, for_IF any_DD text_NN1 ,_, there_EX are_VBR a_AT1 number_NN1 of_IO possible_JJ titles_NN2 ._. 
Correspondingly_RR ,_, we_PPIS2 will_VM suggest_VVI ,_, there_EX is_VBZ ,_, for_IF any_DD text_NN1 ,_, a_AT1 number_NN1 of_IO different_JJ ways_NN2 of_IO expressing_VVG '_GE the_AT topic_NN1 '_GE ._. 
Each_DD1 different_JJ way_NN1 of_IO expressing_VVG '_GE the_AT topic_NN1 '_GE will_NN1 effectively_RR represent_VV0 a_AT1 different_JJ judgement_NN1 of_IO what_DDQ is_VBZ being_VBG written_VVN (_( or_CC talked_VVN )_) about_RP in_II a_AT1 text_NN1 ._. 
As_II an_AT1 illustration_NN1 of_IO this_DD1 point_NN1 ,_, consider_VV0 the_AT text_NN1 in_II (_( 3_MC )_) as_II a_AT1 dusty_JJ fragment_NN1 ,_, recovered_VVN during_II an_AT1 archaeological_JJ dig_NN1 in_II the_AT ruins_NN2 of_IO Minneapolis_NP1 in_II the_AT year_NNT1 2500_MC A.D._RA When_CS asked_VVN what_DDQ the_AT text_NN1 is_58 '_GE about_II '_GE ,_, the_AT discourse_NN1 analyst_NN1 in_II the_AT expedition_NN1 might_VM report_VVI that_CST it_PPH1 is_VBZ about_II '_GE procedures_NN2 used_VVN in_II midtwentieth-century_JJ American_JJ middle-class_JJ culture_NN1 for_IF maintaining_VVG cleanliness_NN1 in_II their_APPGE garments_NN2 '_GE ._. 
(_( Note_VV0 the_AT temporal_JJ and_CC locational_JJ elements_NN2 included_VVD here-elements_NN2 which_DDQ we_PPIS2 shall_VM consider_VVI more_RGR fully_RR later_RRR ._. )_) 
Another_DD1 discourse_NN1 analyst_NN1 ,_, providing_VVG a_AT1 second_MD opinion_NN1 ,_, might_VM report_VVI that_CST it_PPH1 is_VBZ about_II something_PN1 else_RR entirely_RR ,_, and_CC a_AT1 debate_NN1 would_VM ensue_VVI in_II the_AT discourse_NN1 analysis_NN1 literature_NN1 ._. 
The_AT same_DA '_GE text_NN1 '_GE is_VBZ considered_VVN by_II both_DB2 analysts_NN2 ._. 
Their_APPGE disagreement_NN1 would_VM be_VBI over_II ways_NN2 of_IO expressing_VVG '_GE the_AT topic_NN1 '_GE ._. 
(_( Literary_JJ critics_NN2 are_VBR still_RR exercised_VVN about_II the_AT topic_NN1 of_IO Hamlet_NP1 ._. )_) 
The_AT difficulty_NN1 of_IO determining_VVG a_AT1 single_JJ phrase_NN1 or_CC sentence_NN1 as_58 '_GE the_AT topic_NN1 '_GE of_IO a_AT1 piece_NN1 of_IO printed_JJ text_NN1 is_VBZ increased_VVN when_CS fragments_NN2 of_IO conversational_JJ discourse_NN1 are_VBR considered_VVN ._. 
In_II any_DD conversation_NN1 ,_, '_" what_DDQ is_VBZ being_VBG talked_VVN about_II '_" will_VM be_VBI judged_VVN differently_RR at_II different_JJ points_NN2 and_CC the_AT participants_NN2 themselves_PPX2 may_VM not_XX have_VHI identical_JJ views_NN2 of_IO what_DDQ each_DD1 is_VBZ talking_VVG about_II ._. 
People_NN do_VD0 ,_, however_RR ,_, regularly_RR report_VV0 on_II what_DDQ a_AT1 conversation_NN1 was_58 '_GE about_II '_GE ._. 
There_EX are_VBR informal_JJ ways_NN2 of_IO expressing_VVG the_AT topic_NN1 ,_, even_RR in_II conversational_JJ discourse_NN1 ._. 
Topic_NN1 framework_NN1 The_AT discourse_NN1 analyst_NN1 ,_, then_RT ,_, is_VBZ faced_VVN with_IW several_DA2 problems_NN2 when_RRQ he_PPHS1 wishes_VVZ to_TO use_VVI the_AT very_RG attractive_JJ pretheoretical_JJ notion_NN1 of_IO '_GE topic_NN1 '_GE as_58 '_GE what_DDQ is_VBZ being_VBG talked/written_VVN about_II '_GE ._. 
The_AT notion_NN1 is_VBZ attractive_JJ because_CS it_PPH1 seems_VVZ to_TO be_VBI the_AT central_JJ organising_JJ principle_NN1 for_IF a_AT1 lot_NN1 of_IO discourse_NN1 ._. 
It_PPH1 may_VM enable_VVI the_AT analyst_NN1 to_TO explain_VVI why_RRQ several_DA2 sentences_NN2 or_CC utterances_NN2 should_VM be_VBI considered_VVN together_RL as_II a_AT1 set_NN1 of_IO some_DD kind_NN1 ,_, separate_VV0 from_II another_DD1 set_NN1 ._. 
It_PPH1 might_VM also_RR provide_VVI a_AT1 means_NN of_IO distinguishing_JJ fragments_NN2 of_IO discourse_NN1 which_DDQ are_VBR felt_VVN to_TO be_VBI good_JJ ,_, coherent_JJ ,_, examples_NN2 of_IO English_NN1 from_II those_DD2 that_CST are_VBR ,_, intuitively_RR ,_, incoherent_JJ concatenations_NN2 of_IO sentences_NN2 ._. 
Consider_VV0 ,_, for_REX21 example_REX22 ,_, the_AT following_JJ discourse_NN1 fragment_NN1 ,_, taken_VVN from_II Rochester_NP1 &amp;_CC Martin_NP1 (_( 1979_MC :_: 95_MC )_) ._. 
(_( 4_MC )_) Interviewer_NN1 :_: A_AT1 stitch_NN1 in_II time_NNT1 saves_VVZ nine_MC ._. 
What_DDQ does_VDZ that_DD1 mean_VVI ?_? 
Thought-disordered_JJ Speaker_NN1 :_: Oh_UH !_! that_DD1 's_VBZ because_CS all_DB women_NN2 have_VH0 a_AT1 little_JJ bit_NN1 of_IO magic_NN1 to_II them_PPHO2 I_PPIS1 found_VVD that_DD1 out_RP and_CC it_PPH1 's_VHZ called_VVN it_PPH1 's_VBZ sort_RR21 of_RR22 good_JJ magic_NN1 and_CC nine_MC is_VBZ sort_RR21 of_RR22 a_AT1 magic_JJ number_NN1 +_FO like_CS I_PPIS1 've_VH0 got_VVN nine_MC colors_NN2 here_RL you_PPY will_VM notice_VVI I_PPIS1 've_VH0 got_VVN yellow_JJ ,_, green_JJ ,_, blue_JJ ,_, grey_JJ ,_, orange_NN1 ,_, blue_JJ ,_, and_CC navy_NN1 and_CC I_PPIS1 've_VH0 got_VVN black_JJ and_CC I_PPIS1 've_VH0 got_VVN a_AT1 sort_NN1 of_IO clear_JJ white_NN1 the_AT nine_MC colors_NN2 to_II me_PPIO1 they_PPHS2 are_VBR the_AT whole_JJ universe_NN1 and_CC they_PPHS2 symbolize_VV0 every_AT1 man_NN1 ,_, woman_NN1 and_CC child_NN1 in_II the_AT world_NN1 +_FO Rochester_NP1 &amp;_CC Martin_NP1 attempt_VV0 to_TO describe_VVI the_AT connections_NN2 existing_VVG between_II sentences_NN2 in_II discourse_NN1 of_IO this_DD1 type_NN1 ,_, produced_VVN by_II thought-disordered_JJ and_CC schizophrenic_JJ speakers_NN2 ,_, in_II31 terms_II32 of_II33 conceptual_JJ associations_NN2 and_CC lexical_JJ ties_NN2 ._. 
They_PPHS2 point_VV0 out_RP ,_, however_RR ,_, that_DD1 such_DA connections_NN2 are_VBR '_GE unrelated_JJ to_II the_AT conversational_JJ topic_NN1 '_GE ._. 
The_AT notion_NN1 of_IO '_GE topic_NN1 '_GE ,_, though_CS undefined_JJ ,_, seems_VVZ to_TO provide_VVI Rochester_NP1 &amp;_CC Martin_NP1 with_IW a_AT1 natural_JJ criterion_NN1 for_IF distinguishing_VVG between_II the_AT connected_JJ ,_, yet_RR incoherent_JJ ,_, discourse_NN1 of_IO thought-disordered_JJ speakers_NN2 and_CC the_AT coherent_JJ discourse_NN1 of_IO normal_JJ speakers_NN2 ._. 
If_CS there_EX are_VBR ,_, as_CSA we_PPIS2 have_VH0 already_RR argued_VVN ,_, a_AT1 potentially_RR large_JJ number_NN1 of_IO different_JJ ways_NN2 of_IO expressing_VVG '_GE the_AT topic_NN1 '_GE of_IO even_RR a_AT1 short_JJ written_JJ text_NN1 ,_, how_RRQ does_VDZ the_AT analyst_NN1 determine_VVI which_DDQ is_VBZ the_AT one_MC1 correct_JJ expression_NN1 of_IO the_AT topic_NN1 for_IF the_AT text_NN1 ?_? 
One_MC1 answer_NN1 ,_, of_RR21 course_RR22 ,_, is_VBZ to_TO say_VVI that_DD1 ,_, for_IF any_DD practical_JJ purposes_NN2 ,_, there_EX is_VBZ no_AT such_DA thing_NN1 as_II the_AT one_MC1 correct_JJ expression_NN1 of_IO the_AT topic_NN1 for_IF any_DD fragment_NN1 of_IO discourse_NN1 ._. 
There_EX will_VM always_RR be_VBI a_AT1 set_NN1 of_IO possible_JJ expressions_NN2 of_IO the_AT topic_NN1 ._. 
In_II the_AT terms_NN2 used_VVN by_II Tyler_NP1 (_( 1978_MC :_: 452_MC )_) ,_, the_AT '_GE topic_NN1 '_GE can_NN1 only_RR be_VBI '_" one_MC1 possible_JJ paraphrase_NN1 '_GE of_IO a_AT1 sequence_NN1 of_IO utterances_NN2 ._. 
What_DDQ is_VBZ required_VVN is_VBZ a_AT1 characterisation_NN1 of_IO '_GE topic_NN1 '_GE which_DDQ would_VM allow_VVI each_DD1 of_IO the_AT possible_JJ expressions_NN2 ,_, including_II titles_NN2 ,_, to_TO be_VBI considered_VVN (_( partially_RR )_) correct_JJ ,_, thus_RR incorporating_VVG all_DB reasonable_JJ judgements_NN2 of_IO '_" what_DDQ is_VBZ being_VBG talked_VVN about_II '_GE ._. 
We_PPIS2 suggest_VV0 that_CST such_DA a_AT1 characterisation_NN1 can_VM be_VBI developed_VVN in_II31 terms_II32 of_II33 a_AT1 topic_NN1 framework_NN1 ._. 
In_II Chapter_NN1 2_MC ,_, we_PPIS2 discussed_VVD the_AT problem_NN1 for_IF the_AT discourse_NN1 analyst_NN1 of_IO deciding_VVG just_RR what_DDQ features_NN2 of_IO context_NN1 were_VBDR relevant_JJ in_II the_AT interpretation_NN1 of_IO a_AT1 particular_JJ fragment_NN1 of_IO discourse_NN1 ._. 
We_PPIS2 suggested_VVD there_RL that_CST the_AT strategy_NN1 available_JJ to_II him_PPHO1 would_VM be_VBI ,_, on_II the_AT one_MC1 hand_NN1 ,_, to_TO work_VVI predictively_RR in_II31 terms_II32 of_II33 his_APPGE previous_JJ experience_NN1 (_( similar_JJ speakers_NN2 ,_, similar_JJ genres_NN2 ,_, etc._RA )_) and_CC on_II the_AT other_JJ hand_NN1 to_TO examine_VVI the_AT content_NN1 of_IO the_AT text_NN1 ._. 
From_II the_AT content_NN1 of_IO the_AT text_NN1 the_AT analyst_NN1 can_VM ,_, in_II principle_NN1 ,_, determine_VV0 what_DDQ aspects_NN2 of_IO the_AT context_NN1 are_VBR explicitly_RR reflected_VVN in_II the_AT text_NN1 as_II the_AT formal_JJ record_NN1 of_IO the_AT utterance_NN1 ._. 
Those_DD2 aspects_NN2 of_IO the_AT context_NN1 which_DDQ are_VBR directly_RR reflected_VVN in_II the_AT text_NN1 ,_, and_CC which_DDQ need_VV0 to_TO be_VBI called_VVN upon_II to_TO interpret_VVI the_AT text_NN1 ,_, we_PPIS2 shall_VM refer_VVI to_II as_RG activated_JJ features_NN2 of_IO context_NN1 and_CC suggest_VV0 that_CST they_PPHS2 constitute_VV0 the_AT contextual_JJ framework_NN1 within_II which_DDQ the_AT topic_NN1 is_VBZ constituted_VVN ,_, that_REX21 is_REX22 ,_, the_AT topic_NN1 framework_NN1 ._. 
As_II a_AT1 way_NN1 of_IO characterising_VVG the_AT type_NN1 of_IO feature_NN1 which_DDQ will_VM be_VBI required_VVN in_II a_AT1 topic_NN1 framework_NN1 ,_, we_PPIS2 shall_VM examine_VVI a_AT1 fragment_NN1 of_IO conversational_JJ discourse_NN1 and_CC try_VV0 to_TO determine_VVI what_DDQ is_58 '_GE being_VBG talked_VVN about_II '_GE ._. 
The_AT fragment_NN1 ,_, presented_VVN as_CSA (_( 5_MC )_) ,_, is_VBZ not_XX a_AT1 constructed_JJ piece_NN1 of_IO text_NN1 ,_, it_PPH1 is_VBZ taken_VVN from_II a_AT1 recorded_JJ conversation_NN1 ._. 
As_II an_AT1 example_NN1 of_IO discourse_NN1 analysis_NN1 data_NN ,_, it_PPH1 has_VHZ been_VBN selected_VVN for_IF a_AT1 particular_JJ purpose_NN1 ._. 
It_PPH1 is_VBZ not_XX a_AT1 difficult_JJ fragment_NN1 to_TO work_VVI with_IW ,_, it_PPH1 has_VHZ a_AT1 definable_JJ beginning_NN1 and_CC end_NN1 ,_, and_CC ,_, for_IF most_DAT of_IO the_AT fragment_NN1 ,_, there_EX is_VBZ one_MC1 participant_NN1 talking_VVG ,_, in_II31 response_II32 to_II33 another_DD1 's_VBZ request_NN1 for_IF information_NN1 ._. 
This_DD1 request_NN1 for_IF information_NN1 provides_VVZ a_AT1 direction_NN1 for_IF the_AT conversational_JJ fragment_NN1 ,_, so_CS21 that_CS22 we_PPIS2 are_VBR considering_VVG speech_NN1 with_IW some_DD purpose_NN1 and_CC not_XX just_RR social_JJ chat_NN1 used_VMK to_TO pass_VVI the_AT time_NNT1 ._. 
One_PN1 might_VM also_RR say_VVI that_CST the_AT content_NN1 of_IO the_AT request_NN1 for_IF information_NN1 could_VM provide_VVI some_DD basis_NN1 for_IF the_AT content_NN1 of_IO the_AT response_NN1 ,_, especially_RR when_CS the_AT request_NN1 is_VBZ for_IF the_AT meaning_NN1 of_IO an_AT1 expression_NN1 to_TO be_VBI given_VVN ._. 
That_REX21 is_REX22 ,_, it_PPH1 would_VM seem_VVI ,_, at_II first_MD glance_NN1 ,_, to_TO be_VBI a_AT1 simple_JJ matter_NN1 to_TO produce_VVI '_GE the_AT topic_NN1 '_GE for_IF this_DD1 discourse_NN1 fragment_NN1 ,_, for_IF it_PPH1 is_VBZ contained_VVN in_II the_AT question_NN1 asked_VVD ._. 
Immediately_RR prior_II21 to_II22 the_AT following_JJ extract_NN1 ,_, the_AT speaker_NN1 has_VHZ been_VBN asked_VVN the_AT meaning_NN1 of_IO the_AT expression_NN1 ,_, '_GE smoke_NN1 the_AT houses'._NNU (_( 5_MC )_) R_ZZ1 :_: in_II those_DD2 days_NNT2 +_FO when_RRQ we_PPIS2 were_VBDR young_JJ +_FO there_EX was_VBDZ no_AT local_JJ fire_NN1 engine_NN1 here_RL +_FO it_PPH1 was_VBDZ just_RR a_AT1 two-wheeled_JJ trolley_NN1 which_DDQ was_VBDZ kept_VVN in_II the_AT borough_NN1 +_FO in_II the_AT borough_NN1 eh_UH store_VV0 down_RP on_II James_NP1 Street_NNL1 +_FO and_CC whenever_RRQV a_AT1 fire_NN1 broke_VVD out_RP +_FO it_PPH1 was_VBDZ just_RR a_AT1 question_NN1 of_IO whoever_PNQV saw_VVD the_AT fire_NN1 first_MD yelling_NN1 '_GE Fire_NN1 '_GE +_FO and_CC the_AT nearest_JJT people_NN ran_VVD for_IF the_AT trolley_NN1 and_CC how_RRQ they_PPHS2 got_VVD on_RP with_IW it_PPH1 goodness_NN1 knows_VVZ +_FO nobody_PN1 was_VBDZ trained_VVN in_II its_APPGE use_NN1 +_FO anyway_RR everybody_PN1 knew_VVD to_TO go_VVI for_IF the_AT trolley_NN1 +_FO well_RR +_FO when_RRQ we_PPIS2 were_VBDR children_NN2 +_FO we_PPIS2 used_VMK to_TO use_VVI this_DD1 taw_NN1 &lsqb;_( t_ZZ1 :_: &rsqb;_) +_FO it_PPH1 smouldered_VVD furiously_RR +_FO black_JJ thick_JJ smoke_NN1 came_VVD from_II it_PPH1 and_CC we_PPIS2 used_VMK to_TO get_VVI it_PPH1 burning_VVG +_FO and_CC then_RT go_VV0 to_II a_AT1 letter_NN1 box_NN1 and_CC just_RR keep_VV0 blowing_VVG +_FO open_VV0 the_AT letter_NN1 box_NN1 +_FO and_CC just_RR keep_VV0 blowing_VVG the_AT smoke_NN1 in_II +_FO you_PPY see_VV0 +_FO till_CS you_PPY 'd_VM fill_VVI up_RP the_AT lower_JJR part_NN1 of_IO the_AT house_NN1 with_IW nothing_PN1 but_II smoke_NN1 +_FO there_EX was_VBDZ no_AT fire_NN1 +_FO but_CCB just_RR fill_VV0 it_PPH1 up_RP with_IW smoke_NN1 +_FO just_RR to_TO put_VVI the_AT breeze_NN1 up_II +_FO just_RR as_II a_AT1 joke_NN1 +_FO and_CC then_RT of_RR21 course_RR22 +_FO when_RRQ somebody_PN1 would_VM open_VVI a_AT1 window_NN1 or_CC a_AT1 door_NN1 the_AT smoke_NN1 would_VM come_VVI pouring_VVG out_RP +_FO and_CC then_RT +_FO everybody_PN1 was_VBDZ away_RL then_RT for_IF the_AT trolley_NN1 +_FO we_PPIS2 just_RR stood_VVD and_CC watched_VVD all_DB of_IO them_PPHO2 +_FO +_FO S_ZZ1 :_: so_RR that_DD1 's_VBZ what_DDQ '_VBZ smoke_NN1 the_AT houses_NN2 '_GE is_VBZ ?_? 
R_ZZ1 :_: probably_RR +_FO probably_RR +_FO we_PPIS2 called_VVD it_PPH1 '_VBZ the_AT taw_NN1 '_GE +_FO If_CS we_PPIS2 were_VBDR to_TO say_VVI that_CST the_AT topic_NN1 of_IO this_DD1 discourse_NN1 fragment_NN1 is_58 '_GE the_AT meaning_NN1 of_IO the_AT expression_NN1 '_GE smoke_NN1 the_AT houses_NN2 '_GE '_GE ,_, we_PPIS2 could_VM not_XX claim_VVI to_TO have_VHI said_VVN very_RG much_DA1 of_IO analytic_JJ interest_NN1 ._. 
It_PPH1 may_VM be_VBI that_DD1 ,_, for_IF participant_NN1 S_ZZ1 ,_, the_AT above_JJ expression_NN1 represents_VVZ the_AT best_JJT way_NN1 of_IO summarising_VVG what_DDQ speaker_NN1 R_ZZ1 was_VBDZ talking_VVG about_II ,_, as_CSA evidenced_VVN by_II her_APPGE response_NN1 ._. 
However_RR ,_, even_CS21 if_CS22 we_PPIS2 take_VV0 that_DD1 summarising_JJ phrase_NN1 as_CSA one_MC1 possible_JJ expression_NN1 of_IO the_AT topic_NN1 of_IO speaker_NN1 R_ZZ1 's_VBZ lengthy_JJ contribution_NN1 ,_, we_PPIS2 have_VH0 surely_RR not_XX adequately_RR characterised_VVN what_DDQ this_DD1 speaker_NN1 was_VBDZ talking_VVG about_II ._. 
We_PPIS2 might_VM suggest_VVI that_CST the_AT speaker_NN1 is_VBZ talking_VVG about_II a_AT1 joke_NN1 or_CC a_AT1 prank_NN1 ._. 
In_II doing_VDG so_RR ,_, he_PPHS1 talks_VVZ about_II an_AT1 object_NN1 called_VVN '_GE the_AT taw_NN1 '_GE which_DDQ produces_VVZ a_AT1 lot_NN1 of_IO smoke_NN1 ._. 
He_PPHS1 talks_VVZ about_II the_AT process_NN1 of_IO putting_VVG the_AT smoke_NN1 into_II houses_NN2 through_II the_AT letter_NN1 box_NN1 and_CC how_RRQ smoke_NN1 would_VM come_VVI out_II21 of_II22 the_AT window_NN1 or_CC door_NN1 ._. 
He_PPHS1 also_RR talks_VVZ about_II an_AT1 object_NN1 known_VVN as_II the_AT trolley_NN1 ,_, a_AT1 type_NN1 of_IO fire_NN1 engine_NN1 ,_, and_CC the_AT events_NN2 associated_VVN with_IW its_APPGE use_NN1 ._. 
He_PPHS1 talks_VVZ about_II people_NN going_VVG for_IF the_AT trolley_NN1 when_CS the_AT smoke_NN1 comes_VVZ out_II21 of_II22 a_AT1 house_NN1 ._. 
Thus_RR one_MC1 account_NN1 of_IO what_DDQ this_DD1 speaker_NN1 is_VBZ talking_VVG about_II would_VM contain_VVI the_AT following_JJ elements_NN2 :_: a_AT1 joke_NN1 the_AT taw_NN1 smoke_NN1 into_II houses_NN2 out_II21 of_II22 houses_NN2 people_NN get_VV0 trolley_NN1 the_AT use_NN1 of_IO the_AT trolley_NN1 ._. 
This_DD1 set_NN1 of_IO objects_NN2 and_CC events_NN2 could_VM be_VBI taken_VVN as_II a_AT1 set_NN1 of_IO elements_NN2 which_DDQ would_VM have_VHI to_TO be_VBI included_VVN in_II a_AT1 representation_NN1 of_IO this_DD1 speaker_NN1 's_GE topic_NN1 ,_, i.e._REX what_DDQ he_PPHS1 was_VBDZ talking_VVG about_II ._. 
It_PPH1 is_VBZ not_XX a_AT1 complete_JJ set_NN1 ._. 
In_II this_DD1 fragment_NN1 ,_, the_AT speaker_NN1 is_VBZ also_RR talking_NN1 '_GE about_II '_GE a_AT1 particular_JJ time_NNT1 and_CC place_NN1 ,_, and_CC '_GE about_II '_GE a_AT1 specific_JJ person_NN1 ._. 
He_PPHS1 is_VBZ talking_VVG about_II his_APPGE own_DA childhood_NN1 (_( when_CS we_PPIS2 were_VBDR children_NN2 )_) in_II Stornoway_NP1 (_( here_RL )_) ._. 
This_DD1 last_MD element_NN1 presents_VVZ a_AT1 problem_NN1 ,_, because_CS there_EX is_VBZ nothing_PN1 in_II the_AT text_NN1 of_IO the_AT conversational_JJ fragment_NN1 to_TO indicate_VVI this_DD1 location_NN1 ._. 
Yet_RR it_PPH1 is_VBZ a_AT1 piece_NN1 of_IO knowledge_NN1 relevant_JJ to_II what_DDQ the_AT speaker_NN1 is_VBZ talking_VVG about_II and_CC ,_, importantly_RR ,_, knowledge_NN1 which_DDQ the_AT speaker_NN1 assumes_VVZ is_VBZ available_JJ ,_, to_II his_APPGE hearer_NN1 ._. 
Presumably_RR ,_, the_AT speaker_NN1 can_VM also_RR assume_VVI that_DD1 ,_, because_CS his_APPGE hearer_NN1 knows_VVZ ,_, approximately_RR ,_, the_AT speaker_NN1 's_GE age_NN1 ,_, the_AT hearer_NN1 can_VM judge_VVI the_AT time_NNT1 (_( i.e._REX forty_MC years_NNT2 before_CS and_CC not_XX ten_MC years_NNT2 before_II )_) of_IO the_AT events_NN2 described_VVN ._. 
Aspects_NN2 of_IO the_AT speaker_NN1 's_GE assumptions_NN2 about_II his_APPGE hearer_NN1 's_GE knowledge_NN1 must_VM also_RR be_VBI considered_VVN in_II31 relation_II32 to_II33 the_AT elements_NN2 which_DDQ the_AT speaker_NN1 does_VDZ make_VVI explicit_JJ in_II his_APPGE contribution_NN1 ._. 
Do_VD0 the_AT first_MD lines_NN2 of_IO this_DD1 fragment_NN1 contribute_VV0 to_II answering_VVG the_AT question_NN1 asked_VVD ?_? 
Strictly_RR speaking_VVG they_PPHS2 do_VD0 not_XX ._. 
Yet_RR one_PN1 would_VM hesitate_VVI to_TO describe_VVI these_DD2 lines_NN2 as_CSA irrelevant_JJ ._. 
They_PPHS2 are_VBR relevant_JJ to_II what_DDQ the_AT speaker_NN1 wishes_VVZ to_TO provide_VVI as_II an_AT1 answer_NN1 to_II the_AT question_NN1 ,_, given_VVN the_AT particular_JJ hearer_NN1 he_PPHS1 has_VHZ ._. 
This_DD1 young_JJ American_JJ hearer_NN1 ,_, visiting_VVG Stornoway_NP1 ,_, may_VM have_VHI a_AT1 quite_RG inappropriate_JJ idea_NN1 of_IO the_AT type_NN1 of_IO object_NN1 ,_, and_CC the_AT associated_JJ behaviour_NN1 ,_, involved_JJ in_II dealing_VVG with_IW a_AT1 fire_NN1 in_II Stornoway_NP1 forty_MC years_NNT2 before_RT ._. 
Without_IW knowing_VVG about_II the_AT trolley_NN1 ,_, the_AT hearer_NN1 may_VM not_XX (_( in_II the_AT speaker_NN1 's_GE assessment_NN1 perhaps_RR )_) appreciate_VV0 the_AT full_JJ flavour_NN1 of_IO the_AT joke_NN1 or_CC prank_NN1 being_VBG described_VVN ._. 
It_PPH1 may_VM be_VBI argued_VVN that_CST this_DD1 last_MD point_NN1 has_VHZ more_DAR to_TO do_VDI with_IW why_RRQ the_AT speaker_NN1 talked_VVN about_II something_PN1 than_CSN with_IW what_DDQ he_PPHS1 talked_VVD about_II ._. 
Any_DD consideration_NN1 of_IO topic_NN1 involves_VVZ asking_VVG why_RRQ the_AT speaker_NN1 said_VVD what_DDQ he_PPHS1 said_VVD in_II a_AT1 particular_JJ discourse_NN1 situation_NN1 ._. 
As_CSA Coulthard_NP1 (_( 1977_MC :_: 76_MC )_) ,_, following_VVG Sacks_NP1 (_( 1971_MC )_) ,_, points_VVZ out_RP ,_, there_EX is_VBZ a_AT1 constant_JJ analysis_NN1 in_II conversation_NN1 of_IO what_DDQ is_VBZ said_VVN in_II31 terms_II32 of_II33 '_" why_RRQ that_DD1 now_RT and_CC to_II me_PPIO1 '_GE ._. 
In_II the_AT present_JJ discussion_NN1 we_PPIS2 have_VH0 already_RR partially_RR answered_VVN the_AT reader_NN1 's_GE primary_NN1 '_GE why_RRQ '_GE question_NN1 about_II the_AT discourse_NN1 fragment_NN1 being_VBG studied_VVN by_II providing_VVG the_AT previous_JJ speaker_NN1 's_GE question_NN1 ._. 
That_REX21 is_REX22 ,_, attempting_VVG to_TO provide_VVI an_AT1 account_NN1 of_IO what_DDQ a_AT1 person_NN1 is_VBZ talking_VVG about_II is_VBZ always_RR built_VVN on_II an_AT1 assumption_NN1 that_CST we_PPIS2 know_VV0 why_RRQ that_DD1 person_NN1 says_VVZ what_DDQ he_PPHS1 says_VVZ ._. 
The_AT point_NN1 may_VM be_VBI clearer_JJR if_CS we_PPIS2 consider_VV0 a_AT1 possible_JJ reaction_NN1 to_II the_AT expression_NN1 ,_, '_GE Roses_NN2 are_VBR red_JJ ,_, violets_NN2 are_VBR blue_JJ '_" being_VBG included_VVN in_II (_( 5_MC )_) after_II the_AT speaker_NN1 has_VHZ said_VVN nobody_PN1 was_VBDZ trained_VVN in_II its_APPGE use_NN1 ._. 
Would_VM the_AT expression_NN1 simply_RR be_VBI included_VVN in_II the_AT list_NN1 of_IO what_DDQ was_VBDZ talked_VVN about_II ,_, or_CC would_VM it_PPH1 prompt_VVI the_AT question_NN1 '_GE Why_RRQ does_VDZ he_PPHS1 say_VVI that_CST here_RL ?_? 
'_GE The_AT acceptance_NN1 of_IO extract_NN1 (_( 5_MC )_) as_II a_AT1 reasonable_JJ piece_NN1 of_IO English_JJ conversational_JJ discourse_NN1 involves_VVZ implicitly_RR assessing_VVG each_DD1 expression_NN1 in_II31 terms_II32 of_II33 the_AT '_" why_RRQ ?_? '_GE question_NN1 above_RL and_CC finding_VVG a_AT1 suitable_JJ answer_NN1 ._. 
Part_NN1 of_IO the_AT process_NN1 of_IO analysing_VVG discourse_NN1 in_II31 terms_II32 of_II33 '_GE topic_NN1 '_GE is_VBZ an_AT1 attempt_NN1 to_TO make_VVI explicit_JJ the_AT basis_NN1 for_IF our_APPGE intuitive_JJ ability_NN1 to_TO recognise_VVI why_RRQ what_DDQ is_VBZ said_VVN is_VBZ appropriate_JJ in_II a_AT1 particular_JJ discourse_NN1 fragment_NN1 ._. 
Certain_JJ elements_NN2 which_DDQ constrain_VV0 the_AT topic_NN1 can_VM be_VBI determined_VVN before_II this_DD1 discourse_NN1 begins_VVZ ._. 
These_DD2 elements_NN2 are_VBR part_NN1 of_IO what_DDQ ,_, in_II the_AT previous_JJ chapter_NN1 ,_, were_VBDR described_VVN as_II the_AT context_NN1 of_IO a_AT1 speech_NN1 event_NN1 ._. 
In_II relating_VVG contextual_JJ features_NN2 to_II a_AT1 particular_JJ speech_NN1 event_NN1 ,_, however_RR ,_, we_PPIS2 are_VBR particularly_RR interested_JJ in_II only_RR those_DD2 activated_JJ features_NN2 of_IO context_NN1 pertaining_II21 to_II22 the_AT fragment_NN1 of_IO discourse_NN1 being_VBG studied_VVN ._. 
For_REX21 example_REX22 ,_, aspects_NN2 of_IO the_AT time_NNT1 and_CC place_NN1 of_IO the_AT discourse_NN1 in_II (_( 5_MC )_) are_VBR important_JJ because_CS they_PPHS2 have_VH0 a_AT1 bearing_NN1 on_II what_DDQ the_AT speaker_NN1 says_VVZ in_II the_AT fragment_NN1 (_( forty_MC years_NNT2 after_II the_AT described_JJ event_NN1 took_VVD place_NN1 ,_, but_CCB still_RR in_II Stornoway_NP1 )_) ._. 
Similarly_RR ,_, certain_JJ facts_NN2 about_II the_AT speaker_NN1 and_CC hearer_NN1 ,_, as_CSA we_PPIS2 pointed_VVD out_RP earlier_RRR ,_, have_VH0 to_TO be_VBI included_VVN ._. 
As_II a_AT1 first_MD approximation_NN1 ,_, then_RT ,_, we_PPIS2 could_VM produce_VVI a_AT1 partial_JJ representation_NN1 of_IO a_AT1 '_GE framework_NN1 '_GE for_IF extract_NN1 (_( 5_MC )_) in_II31 terms_II32 of_II33 the_AT following_JJ set_NN1 of_IO activated_JJ contextual_JJ features_NN2 ._. 
Conversation_NN1 between_II Participant_NN1 R_ZZ1 (_( 50+_FO years_NNT2 ,_, Scottish_JJ ,_, male_NN1 ,_, ..._... )_) and_CC Participant_NN1 S_ZZ1 (_( 20+_FO years_NNT2 ,_, American_JJ ,_, female_NN1 ,_, ..._... )_) in_II location_NN1 p_ZZ1 (_( Stornoway_NP1 ,_, ..._... )_) at_II time_NNT1 t_ZZ1 (_( late_JJ 1970s_MC2 ,_, ..._... )_) 
This_DD1 simple_JJ set_NN1 of_IO features_NN2 which_DDQ we_PPIS2 have_VH0 claimed_VVN are_VBR necessary_JJ for_IF a_AT1 discussion_NN1 of_IO topic_NN1 are_VBR required_VVN ,_, quite_RG independently_RR of_IO topic_NN1 considerations_NN2 ,_, in_II any_DD form_NN1 of_IO discourse_NN1 analysis_NN1 ._. 
For_IF ethnographers_NN2 and_CC sociolinguists_NN2 considering_II linguistic_JJ interaction_NN1 ,_, these_DD2 elements_NN2 and_CC others_NN2 have_VH0 to_TO be_VBI made_VVN explicit_JJ in_II the_AT analysis_NN1 of_IO features_NN2 such_II21 as_II22 code-switching_NN1 and_CC role-relationships_NN2 ._. 
For_IF the_AT formal_JJ semanticist_NN1 ,_, these_DD2 elements_NN2 are_VBR required_VVN in_II the_AT assignment_NN1 of_IO values_NN2 to_II indexicals_NN2 such_II21 as_II22 I_ZZ1 ,_, you_PPY ,_, here_RL and_CC now_RT ._. 
That_REX21 is_REX22 ,_, in_II building_VVG a_AT1 framework_NN1 for_IF the_AT analysis_NN1 of_IO topic_NN1 ,_, we_PPIS2 are_VBR not_XX adding_VVG any_DD machinery_NN1 to_II the_AT apparatus_NN1 of_IO the_AT discourse_NN1 analyst_NN1 which_DDQ he_PPHS1 does_VDZ not_XX have_VHI to_TO employ_VVI already_RR ._. 
Those_DD2 contextual_JJ features_NN2 we_PPIS2 have_VH0 described_VVN above_RL are_VBR ,_, of_RR21 course_RR22 ,_, derived_VVN from_II the_AT physical_JJ context_NN1 ._. 
They_PPHS2 are_VBR external_JJ to_II the_AT text_NN1 ._. 
There_EX is_VBZ ,_, for_IF most_DAT conversational_JJ fragments_NN2 ,_, a_AT1 set_NN1 of_IO discourseinternal_JJ elements_NN2 which_DDQ are_VBR derived_VVN from_II the_AT conversation_NN1 prior_II21 to_II22 the_AT particular_JJ fragment_NN1 being_VBG studied_VVN ._. 
These_DD2 elements_NN2 are_VBR introduced_VVN in_II the_AT preceding_JJ co-text_NN1 and_CC form_VV0 part_NN1 of_IO what_DDQ has_VHZ been_VBN described_VVN as_II '_GE the_AT domain_NN1 of_IO discourse_NN1 '_GE (_( cf._VV0 Karttunen_NP1 ,_, 1974_MC )_) ._. 
Within_II the_AT domain_NN1 of_IO a_AT1 particular_JJ discourse_NN1 fragment_NN1 are_VBR the_AT people_NN ,_, places_NN2 ,_, entities_NN2 ,_, events_NN2 ,_, facts_NN2 ,_, etc._RA already_RR activated_VVN for_IF both_DB2 participants_NN2 because_CS they_PPHS2 have_VH0 been_VBN mentioned_VVN in_II the_AT preceding_JJ conversation_NN1 ._. 
If_CS the_AT fragment_NN1 of_IO discourse_NN1 one_PN1 wished_VVD to_TO study_VVI was_VBDZ only_RR the_AT part_NN1 of_IO (_( 5_MC )_) beginning_NN1 ,_, when_CS we_PPIS2 were_VBDR children_NN2 we_PPIS2 used_VMK to_TO use_VVI this_DD1 taw_NN1 ,_, then_RT accounting_VVG for_IF the_AT speaker_NN1 's_GE mention_NN1 of_IO the_AT trolley_NN1 near_II the_AT end_NN1 of_IO this_DD1 fragment_NN1 would_VM have_VHI to_TO be_VBI done_VDN in_II31 terms_II32 of_II33 the_AT preceding_JJ discourse_NN1 (_( i.e._REX all_DB the_AT first_MD section_NN1 before_II the_AT taw_NN1 is_VBZ mentioned_VVN )_) in_II which_DDQ the_AT trolley_NN1 is_VBZ introduced_VVN and_CC characterised_VVN ._. 
We_PPIS2 have_VH0 introduced_VVN some_DD basic_JJ components_NN2 which_DDQ would_VM be_VBI required_VVN in_II a_AT1 characterisation_NN1 of_IO the_AT topic_NN1 framework_NN1 for_IF any_DD discourse_NN1 fragment_NN1 ._. 
The_AT topic_NN1 framework_NN1 consists_VVZ of_IO elements_NN2 derivable_JJ from_II the_AT physical_JJ context_NN1 and_CC from_II the_AT discourse_NN1 domain_NN1 of_IO any_DD discourse_NN1 fragment_NN1 ._. 
Notice_VV0 that_CST we_PPIS2 have_VH0 concentrated_VVN on_II only_RR those_DD2 elements_NN2 which_DDQ are_VBR activated_VVN ,_, that_REX21 is_REX22 ,_, relevant_JJ to_II the_AT interpretation_NN1 of_IO what_DDQ is_VBZ said_VVN ._. 
If_CS we_PPIS2 say_VV0 that_CST characterising_VVG the_AT topic_NN1 framework_NN1 is_VBZ a_AT1 means_NN of_IO making_VVG explicit_JJ some_DD of_IO the_AT assumptions_NN2 a_AT1 speaker_NN1 can_VM make_VVI about_II his_APPGE hearer_NN1 's_GE knowledge_NN1 ,_, we_PPIS2 are_VBR not_XX talking_VVG about_II the_AT total_JJ knowledge_NN1 which_DDQ the_AT speaker_NN1 believes_VVZ he_PPHS1 shares_VVZ with_IW his_APPGE hearer_NN1 ._. 
We_PPIS2 are_VBR describing_VVG only_RR that_DD1 activated_JJ part_NN1 which_DDQ is_VBZ required_VVN in_II the_AT analysis_NN1 of_IO the_AT discourse_NN1 fragment_NN1 under_II consideration_NN1 ._. 
This_DD1 approach_NN1 is_VBZ crucially_RR different_JJ from_II some_DD other_JJ proposals_NN2 we_PPIS2 shall_VM examine_VVI ._. 
Presupposition_NN1 pools_NN2 What_DDQ we_PPIS2 have_VH0 described_VVN as_II a_AT1 topic_NN1 framework_NN1 has_VHZ much_RR in_II31 common_II32 with_II33 Venneman_NP1 's_GE proposal_NN1 that_CST ,_, for_IF a_AT1 discourse_NN1 ,_, there_EX is_VBZ a_AT1 presupposition_NN1 pool_NN1 which_DDQ contains_VVZ information_NN1 '_GE constituted_JJ from_II general_JJ knowledge_NN1 ,_, from_II the_AT situative_JJ context_NN1 of_IO the_AT discourse_NN1 ,_, and_CC from_II the_AT completed_JJ part_NN1 of_IO the_AT discourse_NN1 itself_PPX1 '_GE (_( Venneman_NP1 ,_, 1975_MC :_: 314_MC )_) ._. 
In_II this_DD1 approach_NN1 ,_, each_DD1 participant_NN1 in_II a_AT1 discourse_NN1 has_VHZ a_AT1 presupposition_NN1 pool_NN1 and_CC his_APPGE pool_NN1 is_VBZ added_VVN to_II as_II the_AT discourse_NN1 proceeds_NN2 ._. 
Each_DD1 participant_NN1 also_RR behaves_VVZ as_CS21 if_CS22 there_EX exists_VVZ only_RR one_MC1 presupposition_NN1 pool_NN1 shared_VVN by_II all_DB participants_NN2 in_II the_AT discourse_NN1 ._. 
Venneman_NN1 emphasises_VVZ that_CST this_DD1 is_VBZ true_JJ in_II '_GE a_AT1 normal_JJ ,_, honest_JJ discourse_NN1 '_GE ._. 
Within_II the_AT presupposition_NN1 pool_NN1 for_IF any_DD discourse_NN1 ,_, there_EX is_VBZ a_AT1 set_NN1 of_IO discourse_NN1 subjects_NN2 and_CC each_DD1 discourse_NN1 is_VBZ ,_, in_II a_AT1 sense_NN1 ,_, about_II its_APPGE discourse_NN1 subjects_NN2 ._. 
Because_CS it_PPH1 is_VBZ part_NN1 of_IO the_AT shared_JJ assumptions_NN2 of_IO the_AT discourse_NN1 participants_NN2 that_CST these_DD2 discourse_NN1 subjects_NN2 exist_VV0 ,_, they_PPHS2 do_VD0 not_XX need_VVI to_TO have_VHI their_APPGE existence_NN1 asserted_VVN in_II the_AT discourse_NN1 ._. 
Examples_NN2 of_IO expressions_NN2 used_VVN for_IF discourse_NN1 subjects_NN2 might_VM be_VBI the_AT Queen_NN1 ,_, John_NP1 ,_, John_NP1 's_GE wife_NN1 (_( in_II the_AT presupposition_NN1 pool_NN1 by_II virtue_NN1 of_IO general_JJ knowledge_NN1 )_) ,_, your_APPGE hat_NN1 ,_, today_RT (_( from_II the_AT situative_JJ context_NN1 )_) and_CC a_AT1 concert_NN1 of_IO the_AT Berlin_NP1 Philharmonic_JJ 's_GE last_MD year_NNT1 ,_, several_DA2 essays_NN2 (_( from_II the_AT preceding_JJ part_NN1 of_IO the_AT text_NN1 of_IO the_AT discourse_NN1 )_) ._. 
The_AT number_NN1 of_IO discourse_NN1 subjects_NN2 in_II a_AT1 presupposition_NN1 pool_NN1 shared_VVN by_II participants_NN2 in_II a_AT1 discourse_NN1 ,_, particularly_RR participants_NN2 who_PNQS know_VV0 each_PPX221 other_PPX222 quite_RG well_RR ,_, is_VBZ potentially_RR very_RG large_JJ ._. 
How_RRQ does_VDZ the_AT discourse_NN1 analyst_NN1 decide_VV0 which_DDQ discourse_NN1 subjects_NN2 to_TO include_VVI in_II the_AT presupposition_NN1 pool_NN1 for_IF a_AT1 particular_JJ piece_NN1 of_IO conversational_JJ discourse_NN1 ?_? 
Remembering_VVG that_CST any_DD discourse_NN1 data_NN to_II which_DDQ the_AT analyst_NN1 has_VHZ access_NN1 will_VM only_RR be_VBI a_AT1 fragment_NN1 ,_, it_PPH1 would_VM be_VBI extremely_RR difficult_JJ for_IF the_AT analyst_NN1 to_TO predetermine_VVI the_AT complete_JJ set_NN1 of_IO discourse_NN1 subjects_NN2 which_DDQ participants_NN2 share_VV0 prior_II21 to_II22 a_AT1 particular_JJ discourse_NN1 fragment_NN1 ._. 
The_AT most_DAT he_PPHS1 could_VM hope_VVI to_TO provide_VVI would_VM be_VBI a_AT1 partial_JJ set_NN1 ._. 
The_AT problem_NN1 to_TO be_VBI faced_VVN is_VBZ that_DD1 of_IO limiting_VVG the_AT choice_NN1 of_IO the_AT contents_NN2 of_IO even_RR a_AT1 partial_JJ set_NN1 ,_, in_II some_DD non-arbitrary_JJ way_NN1 ._. 
The_AT most_RGT important_JJ principle_NN1 involved_JJ in_II this_DD1 selection_NN1 of_IO Venneman_NP1 's_GE discourse_NN1 subjects_NN2 must_VM have_VHI to_TO do_VDI with_IW their_APPGE relevance_NN1 to_II the_AT particular_JJ discourse_NN1 fragment_NN1 under_II consideration_NN1 ._. 
If_CS ,_, in_II a_AT1 stretch_NN1 of_IO conversational_JJ discourse_NN1 ,_, the_AT participants_NN2 involved_VVN can_VM be_VBI independently_RR known_VVN to_TO have_VHI potential_JJ discourse_NN1 subjects_NN2 such_II21 as_II22 '_GE the_AT Queen_NN1 '_GE ,_, '_GE the_AT Pope_NP1 '_GE ,_, or_CC even_RR '_GE the_AT King_NN1 of_IO Siam_NP1 '_GE ,_, within_II their_APPGE shared_JJ presupposition_NN1 pool_NN1 ,_, but_CCB do_VD0 not_XX mention_VVI the_AT individuals_NN2 ,_, so_RR identified_VVN ,_, in_II their_APPGE conversation_NN1 ,_, it_PPH1 is_VBZ surely_RR unnecessary_JJ to_TO refer_VVI to_II those_DD2 individuals_NN2 in_II the_AT analysis_NN1 of_IO that_DD1 particular_JJ discourse_NN1 fragment_NN1 They_PPHS2 are_VBR ,_, in_II our_APPGE terms_NN2 ,_, not_XX '_GE activated_JJ '_GE ._. 
This_DD1 would_VM lead_VVI to_II the_AT conclusion_NN1 that_CST the_AT relevant_JJ '_GE discourse_NN1 subjects_NN2 '_GE for_IF a_AT1 particular_JJ discourse_NN1 fragment_NN1 must_VM be_VBI those_DD2 to_II which_DDQ reference_NN1 is_VBZ made_VVN in_II the_AT text_NN1 of_IO the_AT discourse_NN1 ._. 
If_CS '_GE mentionedin-the-text_NN1 '_GE is_VBZ taken_VVN as_II the_AT basis_NN1 for_IF selection_NN1 of_IO discourse_NN1 subjects_NN2 ,_, it_PPH1 should_VM be_VBI noted_VVN that_CST the_AT analyst_NN1 is_VBZ ,_, in_II fact_NN1 ,_, attempting_VVG to_TO reconstitute_VVI the_AT presupposition_NN1 pool_NN1 which_DDQ the_AT participants_NN2 must_VM have_VHI had_VHN prior_II21 to_II22 the_AT discourse_NN1 fragment_NN1 being_VBG analysed_VVN ._. 
Such_DA a_AT1 process_NN1 may_VM be_VBI comparable_JJ to_II the_AT experience_NN1 one_PN1 has_VHZ when_CS switching_VVG on_RP the_AT radio_NN1 in_II the_AT middle_NN1 of_IO a_AT1 discussion_NN1 programme_NN1 and_CC trying_VVG to_TO understand_VVI the_AT discussion_NN1 through_II a_AT1 partial_JJ reconstruction_NN1 of_IO what_DDQ must_VM have_VHI been_VBN said_VVN already_RR ,_, who_PNQS the_AT participants_NN2 must_VM be_VBI ,_, and_RR31 so_RR32 on_RR33 ._. 
It_PPH1 does_VDZ suggest_VVI that_CST the_AT only_JJ information_NN1 the_AT discourse_NN1 analyst_NN1 has_VHZ access_NN1 to_TO is_VBZ that_CST contained_VVD in_II the_AT text_NN1 of_IO a_AT1 discourse_NN1 fragment_NN1 ._. 
Sentential_JJ topic_NN1 and_CC the_AT presupposition_NN1 pool_NN1 Of_RR21 course_RR22 ,_, the_AT data_NN for_IF discourse_NN1 analysis_NN1 is_VBZ not_XX limited_VVN to_II anonymous_JJ ,_, decontextualised_JJ texts_NN2 ,_, as_CSA we_PPIS2 have_VH0 argued_VVN already_RR in_II Chapter_NN1 2_MC ._. 
Concentrating_VVG solely_RR on_II the_AT text_NN1 ,_, however_RR ,_, remains_VVZ a_AT1 common_JJ approach_NN1 in_II many_DA2 accounts_NN2 of_IO discourse_NN1 ._. 
It_PPH1 is_VBZ also_RR characteristic_JJ of_IO this_DD1 approach_NN1 that_CST the_AT text_NN1 to_TO be_VBI analysed_VVN is_VBZ constructed_VVN by_II the_AT analyst_NN1 to_TO illustrate_VVI the_AT points_NN2 he_PPHS1 wishes_VVZ to_TO make_VVI ._. 
This_DD1 ,_, unfortunately_RR ,_, is_VBZ the_AT method_NN1 used_VVN by_II Venneman_NP1 who_PNQS ,_, despite_II the_AT promising_JJ breadth_NN1 of_IO analysis_NN1 suggested_VVN by_II the_AT concept_NN1 of_IO a_AT1 presupposition_NN1 pool_NN1 shared_VVN by_II participants_NN2 ,_, restricts_VVZ his_APPGE investigation_NN1 to_II describing_VVG the_AT relationship_NN1 between_II pairs_NN2 of_IO sentences_NN2 ._. 
The_AT notion_NN1 of_IO '_GE topic_NN1 '_GE considered_VVN by_II Venneman_NP1 reflects_VVZ the_AT limitations_NN2 of_IO his_APPGE investigation_NN1 ._. 
He_PPHS1 considers_VVZ the_AT expression_NN1 '_GE topic_NN1 '_GE or_CC '_GE topic_NN1 of_IO a_AT1 discourse_NN1 '_GE as_CSA referring_VVG to_II a_AT1 discourse_NN1 subject_NN1 on_II which_DDQ the_AT attention_NN1 of_IO the_AT participants_NN2 of_IO the_AT discourse_NN1 is_VBZ concentrated_VVN ._. 
Such_DA concentration_NN1 of_IO attention_NN1 is_VBZ usually_RR ,_, though_CS not_XX always_RR ,_, brought_VVN about_RP by_II an_AT1 immediately_RR preceding_VVG textual_JJ mentioning_NN1 of_IO the_AT discourse_NN1 subject_NN1 ._. 
(_( Venneman_NN1 ,_, 1975_MC 317_MC )_) This_DD1 definition_NN1 of_IO topic_NN1 has_VHZ a_AT1 certain_JJ intuitive_JJ appeal_NN1 ,_, in_II the_AT sense_NN1 that_CST what_DDQ two_MC participants_NN2 are_VBR concentrating_VVG on_RP ,_, in_II their_APPGE conversational_JJ talk_NN1 for_REX21 example_REX22 ,_, is_VBZ a_AT1 reasonable_JJ candidate_NN1 for_IF '_GE the_AT topic_NN1 '_GE ._. 
There_EX are_VBR ,_, however_RR ,_, two_MC basic_JJ problems_NN2 here_RL ._. 
First_MD ,_, this_DD1 definition_NN1 of_IO topic_NN1 seems_VVZ to_TO be_VBI based_VVN on_II the_AT same_DA '_GE topic_NN1 =_FO single_JJ term_NN1 title_NN1 '_GE notion_NN1 which_DDQ we_PPIS2 challenged_VVD earlier_RRR ._. 
As_CSA we_PPIS2 pointed_VVD out_RP then_RT ,_, although_CS a_AT1 stretch_NN1 of_IO discourse_NN1 can_VM appear_VVI to_TO be_VBI largely_RR concerned_JJ with_IW a_AT1 single_JJ individual_NN1 ,_, or_CC one_MC1 discourse_NN1 subject_NN1 ,_, so_CS21 that_CS22 the_AT discourse_NN1 may_VM be_VBI loosely_RR reported_VVN as_CSA being_NN1 '_GE about_II '_" that_DD1 individual_NN1 ,_, this_DD1 should_VM not_XX lead_VVI us_PPIO2 to_TO claim_VVI that_CST all_DB discourses_NN2 are_VBR about_II single_JJ individuals_NN2 or_CC can_VM be_VBI given_VVN convenient_JJ one-word_JJ titles_NN2 ._. 
A_AT1 second_MD objection_NN1 is_VBZ that_CST it_PPH1 is_VBZ far_RG21 from_RG22 clear_JJ how_RRQ we_PPIS2 would_VM decide_VVI ,_, in_II any_DD principled_JJ way_NN1 ,_, what_DDQ the_AT participants_NN2 in_II a_AT1 discourse_NN1 fragment_NN1 are_VBR ,_, in_II fact_NN1 ,_, '_GE concentrating_NN1 '_GE on_RP ._. 
An_AT1 attempt_NN1 is_VBZ made_VVN by_II Venneman_NP1 to_TO provide_VVI a_AT1 formal_JJ means_NN of_IO identifying_VVG the_AT topic_NN1 in_II a_AT1 discourse_NN1 fragment_NN1 ._. 
He_PPHS1 suggests_VVZ that_DD1 like_NN1 '_GE all_DB phenomena_NN2 whose_DDQGE unique_JJ existence_NN1 is_VBZ presupposed_VVN ,_, topics_NN2 can_VM be_VBI referred_VVN to_II by_II means_NN of_IO individual_JJ names_NN2 ,_, deictic_JJ expressions_NN2 ,_, and_CC definite_JJ descriptions_NN2 '_GE (_( Venneman_NP1 ,_, 1975_MC :_: 317_MC )_) ._. 
Using_VVG this_DD1 guide_NN1 ,_, the_AT analyst_NN1 must_VM find_VVI that_CST the_AT following_NN1 two_MC discourse_NN1 fragments_NN2 ,_, one_MC1 each_DD1 from_II stretches_NN2 of_IO spoken_JJ and_CC written_JJ discourse_NN1 ,_, have_VH0 several_DA2 such_DA '_GE topics'._NNU (_( 6_MC )_) what_DDQ was_VBDZ interesting_JJ was_VBDZ that_DD1 little_JJ Richard_NP1 came_VVD home_RL from_II his_APPGE Toronto_NP1 school_NN1 with_IW his_APPGE Newfie_NN1 jokes_VVZ the_AT content_NN1 of_IO which_DDQ the_AT substantive_JJ content_NN1 was_VBDZ identical_JJ to_II Irish_JJ jokes_NN2 which_DDQ my_APPGE son_NN1 comes_VVZ home_RL with_IW from_II Edinburgh_NP1 schools_NN2 (_( 7_MC )_) so_RR can_VM he_PPHS1 ,_, but_CCB the_AT main_JJ point_NN1 about_II this_DD1 system_NN1 is_VBZ the_AT strain_NN1 it_PPH1 puts_VVZ on_RP the_AT other_JJ players_NN2 What_DDQ is_58 '_GE the_AT topic_NN1 '_GE of_IO (_( 6_MC )_) little_JJ Richard_NP1 or_CC his_APPGE Toronto_NP1 school_NN1 or_CC his_APPGE Newfie_NN1 jokes_NN2 ,_, etc._RA ;_; and_CC is_VBZ he_PPHS1 ,_, this_DD1 system_NN1 or_CC the_AT other_JJ players_NN2 the_AT topic_NN1 of_IO (_( 7_MC )_) ?_? 
It_PPH1 is_VBZ possible_JJ to_TO make_VVI a_AT1 guess_NN1 at_II what_DDQ the_AT speaker_NN1 of_IO (_( 6_MC )_) and_CC the_AT writer_NN1 of_IO (_( 7_MC )_) were_VBDR concentrating_VVG on_RP ,_, but_CCB the_AT guess_NN1 is_VBZ probably_RR based_VVN on_II an_AT1 elaborate_JJ reconstruction_NN1 of_IO what_DDQ the_AT most_RGT probable_JJ context_NN1 was_VBDZ ,_, both_RR verbal_JJ and_CC non-verbal_JJ ,_, for_IF these_DD2 two_MC discourse_NN1 fragments_NN2 ._. 
That_REX21 is_REX22 ,_, the_AT reader_NN1 will_VM be_VBI forced_VVN to_TO use_VVI these_DD2 '_GE texts_NN2 '_GE to_TO reconstruct_VVI ,_, not_XX just_RR some_DD relevant_JJ discourse_NN1 subjects_NN2 in_II the_AT presupposition_NN1 pool_NN1 ,_, following_VVG Venneman_NP1 ,_, but_CCB rather_RR some_DD of_IO the_AT elements_NN2 of_IO the_AT topic_NN1 framework_NN1 existing_JJ when_CS these_DD2 discourse_NN1 fragments_NN2 were_VBDR produced_VVN ._. 
It_PPH1 is_VBZ also_RR likely_JJ that_CST the_AT reader_NN1 ,_, if_CS asked_VVN to_TO give_VVI the_AT topic_NN1 for_IF each_DD1 fragment_NN1 ,_, would_VM not_XX simply_RR produce_VVI a_AT1 single-term_NN1 '_GE title_NN1 '_GE ._. 
If_CS the_AT same_DA reader_NN1 were_VBDR faced_VVN with_IW the_AT type_NN1 of_IO '_GE discourse_NN1 '_GE fragment_NN1 created_VVN by_II Venneman_NP1 ,_, reproduced_VVN as_CSA (_( 8_MC )_) below_RL ,_, he_PPHS1 might_VM quite_RG readily_RR provide_VVI support_NN1 for_IF Venneman_NP1 's_GE analysis_NN1 by_II saying_VVG that_DD1 '_VBZ the_AT topic_NN1 '_GE is_VBZ Mary._NP1 (_( 8_MC )_) Mary_NP1 is_VBZ singing_VVG strangely_RR ._. 
The_AT reader_NN1 presumably_RR can_VM just_RR as_RG easily_RR reconstruct_VVI an_AT1 alternative_JJ context_NN1 (_( e.g._REX a_AT1 description_NN1 of_IO the_AT effects_NN2 of_IO marijuana_NN1 on_II a_AT1 Nativity_NN1 play_NN1 performance_NN1 )_) in_II which_DDQ Mary_NP1 would_VM not_XX be_VBI proposed_VVN as_58 '_GE the_AT topic_NN1 of_IO the_AT discourse_NN1 '_GE ._. 
Thus_RR ,_, while_CS there_EX may_VM be_VBI preferences_NN2 discernible_JJ in_II the_AT choice_NN1 of_IO elements_NN2 within_II a_AT1 sentence_NN1 if_CS that_DD1 sentence_NN1 is_VBZ presented_VVN in_II isolation_NN1 ,_, such_DA preferences_NN2 may_VM reflect_VVI the_AT rather_RG trivial_JJ fact_NN1 that_CST names_NN2 are_VBR more_RGR salient_JJ than_CSN anything_PN1 else_RR ,_, in_II isolation_NN1 ._. 
That_CST these_DD2 preferences_NN2 do_VD0 have_VHI significance_NN1 for_IF an_AT1 analysis_NN1 of_IO the_AT syntactic_JJ structure_NN1 of_IO sentences_NN2 has_VHZ been_VBN argued_VVN by_II Kuno_NP1 &amp;_CC Kaburaki_NP1 (_( 1977_MC )_) ._. 
However_RR ,_, it_PPH1 is_VBZ ,_, in_II principle_NN1 ,_, impossible_JJ for_IF a_AT1 discourse_NN1 to_TO consist_VVI of_IO a_AT1 single_JJ decontextualised_JJ sentence_NN1 and_CC ,_, in_II practice_NN1 ,_, rare_JJ for_IF discourse_NN1 participants_NN2 to_TO have_VHI to_TO work_VVI out_RP '_GE the_AT topic_NN1 of_IO discourse_NN1 '_GE one_MC1 sentence_NN1 at_II a_AT1 time_NNT1 ._. 
The_AT most_RRT a_AT1 discourse_NN1 analyst_NN1 could_VM say_VVI about_II a_AT1 discourse_NN1 fragment_NN1 such_II21 as_II22 the_AT sentence_NN1 in_II (_( 8_MC )_) above_RL is_VBZ that_CST Mary_NP1 is_VBZ potentially_RR part_NN1 of_IO the_AT topic_NN1 of_IO the_AT discourse_NN1 in_II which_DDQ (_( 8_MC )_) occurred_VVD ,_, but_CCB more_DAR information_NN1 is_VBZ required_VVN ,_, as_CSA indeed_RR is_VBZ also_RR the_AT case_NN1 for_IF both_DB2 extracts_NN2 (_( 6_MC )_) and_CC (_( 7_MC )_) ._. 
It_PPH1 should_VM be_VBI apparent_JJ that_CST the_AT use_NN1 of_IO single_JJ constructed_JJ sentences_NN2 as_II the_AT basis_NN1 for_IF making_VVG claims_NN2 about_II notions_NN2 such_II21 as_II22 '_GE the_AT topic_NN1 of_IO a_AT1 discourse_NN1 '_GE is_VBZ extremely_RR misleading_JJ ._. 
Relevance_NN1 and_CC speaking_VVG topically_RR The_AT topic_NN1 framework_NN1 ,_, as_CSA we_PPIS2 have_VH0 described_VVN it_PPH1 ,_, represents_VVZ the_AT area_NN1 of_IO overlap_NN1 in_II the_AT knowledge_NN1 which_DDQ has_VHZ been_VBN activated_VVN and_CC is_VBZ shared_VVN by_II the_AT participants_NN2 at_II a_AT1 particular_JJ point_NN1 in_II a_AT1 discourse_NN1 ._. 
Once_CS the_AT elements_NN2 in_II the_AT topic_NN1 framework_NN1 and_CC the_AT interrelationships_NN2 between_II them_PPHO2 have_VH0 been_VBN identified_VVN ,_, the_AT analyst_NN1 has_VHZ some_DD basis_NN1 for_IF making_VVG judgements_NN2 of_IO relevance_NN1 with_II31 regard_II32 to_II33 conversational_JJ contributions_NN2 ._. 
The_AT technical_JJ use_NN1 of_IO the_AT term_NN1 '_GE relevance_NN1 '_GE in_II the_AT analysis_NN1 of_IO conversation_NN1 is_VBZ derived_VVN from_II the_AT conversational_JJ maxims_NN2 proposed_VVN by_II Grice_NP1 (_( 1975_MC )_) ._. 
If_CS ,_, as_CSA Grice_NP1 suggests_VVZ ,_, there_EX is_VBZ a_AT1 general_JJ agreement_NN1 of_IO co-operation_NN1 between_II participants_NN2 in_II conversation_NN1 ,_, then_RT each_DD1 participant_NN1 can_VM expect_VVI the_AT other_JJ to_TO conform_VVI to_II certain_JJ conventions_NN2 in_II speaking_VVG ._. 
These_DD2 conventions_NN2 or_CC maxims_NN2 have_VH0 to_TO do_VDI with_IW the_AT quantity_NN1 (_( or_CC informativeness_NN1 )_) ,_, the_AT quality_NN1 (_( truthfulness_NN1 )_) ,_, the_AT manner_NN1 (_( clearness_NN1 )_) and_CC relevance_NN1 of_IO conversational_JJ contributions_NN2 ._. 
Although_CS he_PPHS1 discusses_VVZ and_CC exemplifies_VVZ the_AT other_JJ maxims_NN2 ,_, Grice_NP1 does_VDZ not_XX elaborate_VVI on_II the_AT simple_JJ instruction_NN1 '_GE Be_VBI relevant_JJ ._. 
'_GE The_AT discourse_NN1 analyst_NN1 wishing_VVG to_TO make_VVI use_NN1 of_IO this_DD1 notion_NN1 is_VBZ immediately_RR confronted_VVN with_IW the_AT problem_NN1 of_IO deciding_JJ '_GE relevant_JJ to_II what_DDQ ?_? 
'_" One_MC1 way_NN1 of_IO solving_VVG this_DD1 problem_NN1 is_VBZ to_TO translate_VVI the_AT maxim_NN1 '_GE Be_VBI relevant_JJ '_GE into_II a_AT1 more_RGR practically_RR useful_JJ form_NN1 as_58 '_GE Make_VV0 your_APPGE contribution_NN1 relevant_JJ in_II31 terms_II32 of_II33 the_AT existing_JJ topic_NN1 framework_NN1 ._. 
'_" What_DDQ we_PPIS2 have_VH0 characterised_VVN as_II a_AT1 convention_NN1 of_IO conversational_JJ discourse_NN1 '_GE making_VVG your_APPGE contribution_NN1 relevant_JJ in_II31 terms_II32 of_II33 the_AT existing_JJ topic_NN1 framework_NN1 '_GE could_VM be_VBI captured_VVN more_RGR succinctly_RR in_II the_AT expression_NN1 speaking_VVG topically_RR ._. 
We_PPIS2 could_VM say_VVI that_CST a_AT1 discourse_NN1 participant_NN1 is_58 '_GE speaking_VVG topically_RR '_GE when_CS he_PPHS1 makes_VVZ his_APPGE contribution_NN1 fit_VVI closely_RR to_II the_AT most_RGT recent_JJ elements_NN2 incorporated_VVN in_II the_AT topic_NN1 framework_NN1 ._. 
This_DD1 is_VBZ most_RGT noticeable_JJ in_II conversations_NN2 where_CS each_DD1 participant_NN1 '_GE picks_VVZ up_RP '_GE elements_NN2 from_II the_AT contribution_NN1 of_IO the_AT preceding_JJ speaker_NN1 and_CC incorporates_VVZ them_PPHO2 in_II his_APPGE contribution_NN1 ,_, as_CSA in_II the_AT following_JJ fragment_NN1 :_: (_( 9_MC )_) E_ZZ1 :_: I_PPIS1 went_VVD to_II Yosemite_NP1 National_JJ Park_NN1 F_ZZ1 :_: did_VDD you_PPY E_NP1 :_: yeah_UH it_PPH1 's_VBZ beautiful_JJ there_RL right_RR throughout_II the_AT year_NNT1 +_FO F_ZZ1 :_: I_PPIS1 have_VH0 relations_NN2 in_II California_NP1 and_CC that_DD1 's_VBZ their_APPGE favourite_JJ Park_NN1 because_CS they_PPHS2 +_FO enjoy_VV0 camping_VVG a_AT1 lot_NN1 E_ZZ1 :_: oh_UH yeah_UH F_NP1 :_: they_PPHS2 go_VV0 round_JJ camping_NN1 +_FO E_ZZ1 :_: I_PPIS1 must_VM admit_VVI I_PPIS1 hate_VV0 camping_NN1 +_FO This_DD1 type_NN1 of_IO '_GE speaking_VVG topically_RR '_GE is_VBZ an_AT1 obvious_JJ feature_NN1 of_IO casual_JJ conversation_NN1 in_II which_DDQ each_DD1 participant_NN1 contributes_VVZ equally_RR and_CC there_EX is_VBZ no_AT fixed_JJ direction_NN1 for_IF the_AT conversation_NN1 to_TO go_VVI ._. 
In_II contrast_NN1 ,_, there_EX is_VBZ the_AT type_NN1 of_IO conversational_JJ situation_NN1 in_II which_DDQ the_AT participants_NN2 are_VBR concentrating_VVG their_APPGE talk_NN1 on_II one_MC1 particular_JJ entity_NN1 ,_, individual_JJ or_CC issue_NN1 ._. 
In_II such_DA a_AT1 situation_NN1 ,_, the_AT participants_NN2 may_VM ,_, in_II fact_NN1 ,_, '_GE speak_VV0 topically_RR '_GE ,_, but_CCB they_PPHS2 might_VM also_RR be_VBI said_VVN to_TO be_VBI speaking_VVG on_II a_AT1 topic_NN1 ._. 
An_AT1 extreme_JJ example_NN1 of_IO '_GE speaking_NN1 on_II a_AT1 topic_NN1 '_GE would_VM be_VBI in_II a_AT1 debate_NN1 where_RRQ one_MC1 participant_NN1 ignored_VVD the_AT previous_JJ speaker_NN1 's_GE contribution_NN1 on_II '_GE capital_NN1 punishment_NN1 '_GE ,_, for_REX21 example_REX22 ,_, and_CC presented_VVD his_APPGE talk_NN1 quite_RG independently_RR of_IO any_DD connection_NN1 with_IW what_DDQ went_VVD before_RT ._. 
In_II practice_NN1 ,_, we_PPIS2 should_VM find_VVI that_CST any_DD conversational_JJ fragment_NN1 will_VM exhibit_VVI patterns_NN2 of_IO talk_NN1 in_II which_DDQ both_RR '_GE speaking_VVG topically_RR '_GE and_CC '_GE speaking_NN1 on_II a_AT1 topic_NN1 '_GE are_VBR present_JJ ._. 
Both_DB2 forms_NN2 are_VBR based_VVN on_II the_AT existing_JJ topic_NN1 framework_NN1 ,_, but_CCB the_AT distinction_NN1 derives_VVZ from_II what_DDQ each_DD1 individual_JJ speaker_NN1 treats_VVZ as_II the_AT salient_JJ elements_NN2 in_II the_AT existing_JJ topic_NN1 framework_NN1 ._. 
It_PPH1 is_VBZ quite_RG often_RR the_AT case_NN1 that_CST a_AT1 speaker_NN1 will_VM treat_VVI what_DDQ he_PPHS1 was_VBDZ talking_VVG about_II in_II his_APPGE last_MD contribution_NN1 as_II the_AT most_RGT salient_JJ elements_NN2 and_CC what_DDQ the_AT other_JJ speaker_NN1 talked_VVN about_II ,_, though_CS more_RGR recent_JJ ,_, as_RG less_RGR salient_JJ ._. 
This_DD1 facet_NN1 of_IO conversational_JJ discourse_NN1 quite_RG naturally_RR leads_VVZ to_II a_AT1 consideration_NN1 of_IO the_AT individual_JJ speaker_NN1 's_GE topics_NN2 within_II what_DDQ we_PPIS2 have_VH0 been_VBN discussing_VVG as_II the_AT conversational_JJ topic_NN1 ._. 
Before_CS we_PPIS2 explore_VV0 the_AT influence_NN1 of_IO '_GE speaker_NN1 's_GE topic_NN1 '_GE ,_, we_PPIS2 shall_VM try_VVI to_TO illustrate_VVI in_II some_DD detail_NN1 the_AT way_NN1 in_II which_DDQ conversational_JJ participants_NN2 '_GE speak_VV0 topically_RR '_GE ,_, by_II making_VVG their_APPGE contributions_NN2 relevant_JJ to_II the_AT existing_JJ topic_NN1 framework_NN1 ._. 
In_II the_AT representation_NN1 of_IO the_AT topic_NN1 framework_NN1 ,_, we_PPIS2 shall_VM present_VVI the_AT elements_NN2 involved_VVN as_II a_AT1 list_NN1 ._. 
It_PPH1 is_VBZ difficult_JJ to_TO imagine_VVI an_AT1 appropriate_JJ '_GE diagram_NN1 '_GE which_DDQ could_VM incorporate_VVI both_DB2 the_AT sequential_JJ pattern_NN1 of_IO elements_NN2 introduced_VVN and_CC the_AT interrelatedness_NN1 of_IO those_DD2 elements_NN2 with_IW each_PPX221 other_PPX222 and_CC with_IW the_AT contextual_JJ features_NN2 ._. 
For_IF the_AT moment_NN1 ,_, we_PPIS2 shall_VM identify_VVI some_DD of_IO the_AT elements_NN2 and_CC links_NN2 which_DDQ are_VBR pertinent_JJ to_II an_AT1 analysis_NN1 of_IO one_MC1 fragment._NNU (_( 10_MC )_) Partial_JJ topic_NN1 framework_NN1 existing_VVG in_II a_AT1 conversation_NN1 between_II K_ZZ1 (_( 20+_FO ,_, female_NN1 ,_, Edinburgh-resident_NN1 ,_, university_NN1 student_NN1 ,_, ..._... )_) and_CC J_ZZ1 (_( 60+_FO ,_, male_NN1 ,_, Edinburgh-resident_NN1 ,_, retired_VVD ,_, ..._... )_) in_II P_ZZ1 Working_JJ Men_NN2 's_GE Club_NN1 ,_, Edinburgh_NP1 ,_, ..._... )_) at_II T_ZZ1 (_( early_JJ evening_NNT1 ,_, spring_NN1 ,_, 1976_MC ,_, ..._... )_) mentioning_VVG (_( J_ZZ221 's_ZZ222 three_MC children_NN2 J_ZZ1 '_GE s_ZZ1 brothers_NN2 the_AT schools_NN2 they_PPHS2 attended_VVD the_AT schools_NN2 J_ZZ1 attended_VVD that_CST J_NP1 did_VDD badly_RR at_II school_NN1 J_ZZ1 left_JJ school_NN1 at_II fourteen_MC )_) when_RRQ K_ZZ1 asks_VVZ J_ZZ1 what_DDQ he_PPHS1 did_VDD after_CS he_PPHS1 left_VVD school_NN1 J_ZZ1 :_: oh_UH I_PPIS1 done_VDN odd_JJ jobs_NN2 like_II +_FO paper_NN1 boy_NN1 +_FO chemist_NN1 's_GE shop_NN1 worked_VVN in_II a_AT1 chemist_NN1 shop_NN1 +_FO and_CC done_VDN two_MC or_CC three_MC others+_FO and_CC I_PPIS1 finally_RR started_VVD in_II the_AT bricklaying_NN1 +_FO so_CS I_PPIS1 served_VVD my_APPGE time_NNT1 as_II a_AT1 bricklayer_NN1 +_FO K_ZZ1 :_: that_DD1 's_VBZ good_JJ money_NN1 J_ZZ1 :_: nowadays_RT it_PPH1 is_VBZ but_CCB in_CS21 that_CS22 +_FO when_RRQ my_APPGE time_NNT1 was_VBDZ out_RP it_PPH1 wasn't+_FO it_PPH1 was_VBDZ only_RR three_MC pounds_NN2 nine_MC a_AT1 week_NNT1 +_FO so_CS +_FO +_FO K_ZZ1 :_: my_APPGE father_NN1 was_VBDZ a_AT1 stonemason_NN1 and_CC he_PPHS1 started_VVD at_II home_NN1 +_FO and_CC they_PPHS2 were_VBDR paid_VVN a_AT1 halfpenny_NN1 an_AT1 hour_NNT1 extra_JJ for_IF being_VBG left-handed_RR +_FO +_FO Given_VVN a_AT1 fragment_NN1 of_IO conversation_NN1 and_CC a_AT1 topic_NN1 framework_NN1 as_CSA in_II (_( 10_MC )_) ,_, it_PPH1 is_VBZ possible_JJ for_IF the_AT analyst_NN1 to_TO point_VVI out_RP some_DD ways_NN2 in_II which_DDQ each_DD1 participant_NN1 '_GE speaks_VVZ topically_RR '_GE ._. 
Such_DA an_AT1 undertaking_NN1 can_VM appear_VVI to_TO be_VBI a_AT1 matter_NN1 of_IO stating_VVG the_AT obvious_JJ that_CST speaker_NN1 J_ZZ1 ,_, in_II his_APPGE first_MD contribution_NN1 ,_, for_REX21 example_REX22 ,_, is_VBZ answering_VVG the_AT '_" what_DDQ '_VBZ question_NN1 in_II31 terms_II32 of_II33 an_AT1 understood-to-be-known_JJ location_NN1 and_CC a_AT1 time_NNT1 which_DDQ is_VBZ known_VVN from_II an_AT1 interaction_NN1 between_II knowledge_NN1 of_IO J_ZZ221 's_ZZ222 age_VV0 (_( context_NN1 )_) and_CC knowledge_NN1 that_CST J_ZZ1 was_VBDZ at_RR21 least_RR22 fourteen_MC (_( domain_NN1 )_) ._. 
We_PPIS2 might_VM highlight_VVI the_AT '_GE topicality_NN1 '_GE or_CC '_GE relevance_NN1 '_GE of_IO J_ZZ221 's_ZZ222 first_MD contribution_NN1 by_II asking_VVG how_RRQ K_ZZ1 might_VM have_VHI reacted_VVN if_CS J_ZZ1 had_VHD talked_VVN about_II one_MC1 of_IO his_APPGE brothers_NN2 ,_, or_CC about_II the_AT type_NN1 of_IO work_NN1 to_TO be_VBI had_VHN in_II Australia_NP1 ,_, or_CC training_NN1 to_TO be_VBI a_AT1 brain_NN1 surgeon_NN1 ._. 
Given_VVN this_DD1 topic_NN1 framework_NN1 ,_, J_ZZ1 is_VBZ constrained_VVN from_II talking_VVG about_II these_DD2 things_NN2 unless_CS he_PPHS1 introduces_VVZ into_II the_AT topic_NN1 framework_NN1 some_DD additional_JJ information_NN1 which_DDQ he_PPHS1 could_VM then_RT treat_VVI as_CSA shared_VVN by_II his_APPGE hearer-that_RG one_MC1 of_IO his_APPGE brothers_NN2 had_VHD gone_VVN to_II Australia_NP1 to_TO train_VVI as_II a_AT1 brain_NN1 surgeon_NN1 and_CC he_PPHS1 considered_VVD doing_VDG the_AT same_DA ,_, but_CCB settled_VVD for_IF bricklaying_NN1 instead_RR ._. 
Thus_RR ,_, J_ZZ221 's_ZZ222 first_MD contribution_NN1 here_RL can_VM be_VBI judged_VVN to_TO be_VBI relevant_JJ in_II31 terms_II32 of_II33 the_AT existing_JJ topic_NN1 framework_NN1 and_CC also_RR to_TO add_VVI some_DD information_NN1 to_II the_AT topic_NN1 framework_NN1 ._. 
In_II this_DD1 first_MD contribution_NN1 ,_, he_PPHS1 does_VDZ n't_XX talk_VVI about_II '_" being_VBG fourteen_MC or_CC older_JJR '_GE or_CC '_GE Edinburgh_NP1 '_GE ,_, but_CCB he_PPHS1 does_VDZ talk_VVI about_II '_GE starting_NN1 work_NN1 as_II a_AT1 bricklayer_NN1 '_GE (_( when_CS I_PPIS1 was_VBDZ fourteen_MC or_CC older_JJR ,_, in_II Edinburgh_NP1 )_) and_CC ,_, as_CSA a_AT1 co-operative_JJ conversationalist_NN1 ,_, he_PPHS1 would_VM have_VHI to_TO state_VVI explicitly_RR if_CS the_AT information_NN1 '_GE being_NN1 fourteen_MC or_CC older_JJR ,_, in_II Edinburgh_NP1 '_GE was_VBDZ not_XX applicable_JJ ._. 
More_RGR interesting_JJ is_VBZ speaker_NN1 K_ZZ221 's_ZZ222 first_MD contribution_NN1 in_II (_( 10_MC )_) ._. 
First_MD ,_, its_APPGE connection_NN1 to_II the_AT preceding_JJ discourse_NN1 depends_VVZ on_II a_AT1 general_JJ inference_NN1 that_CST if_CS one_PN1 works_VVZ (_( e.g._REX as_II a_AT1 bricklayer_NN1 )_) one_PN1 receives_VVZ money_NN1 ._. 
(_( We_PPIS2 shall_VM discuss_VVI the_AT role_NN1 of_IO inference_NN1 in_II discourse_NN1 in_II Chapter_NN1 7_MC ._. )_) 
Second_MD ,_, this_DD1 contribution_NN1 has_VHZ the_AT potential_JJ to_TO produce_VVI some_DD conflict_NN1 within_II the_AT conversation_NN1 ,_, since_RR '_GE what_DDQ is_VBZ being_VBG talked_VVN about_II '_GE up_II21 to_II22 this_DD1 point_NN1 is_VBZ not_XX present_JJ time_NNT1 ._. 
The_AT speaker_NN1 appears_VVZ to_TO be_VBI generalising_VVG to_II a_AT1 time_NNT1 which_DDQ includes_VVZ her_APPGE own_DA experience_NN1 ._. 
Within_II the_AT existing_JJ topic_NN1 framework_NN1 ,_, speaker_NN1 K_ZZ1 's_VBZ saying_VVG that_DD1 's_VBZ good_JJ money_NN1 is_VBZ an_AT1 example_NN1 of_IO speaking_VVG topically_RR ,_, for_IF her_PPHO1 ,_, but_CCB ,_, for_IF speaker_NN1 J_ZZ1 ,_, the_AT time_NNT1 co-ordinate_VV0 within_II the_AT topic_NN1 framework_NN1 has_VHZ been_VBN narrowed_VVN down_RP by_II his_APPGE preceding_JJ remarks_NN2 ._. 
There_EX is_VBZ ,_, then_RT ,_, a_AT1 discrepancy_NN1 between_II what_DDQ each_DD1 participant_NN1 is_VBZ talking_VVG about_II ,_, within_II the_AT topic_NN1 framework_NN1 ._. 
We_PPIS2 shall_VM examine_VVI this_DD1 effect_NN1 of_IO individual_JJ speaker_NN1 's_GE topics_NN2 in_II the_AT next_MD section_NN1 ._. 
Speaker_NN1 J_ZZ1 relates_VVZ his_APPGE subsequent_JJ remarks_NN2 to_II the_AT two_MC salient_JJ time_NNT1 co-ordinates_VVZ within_II the_AT topic_NN1 framework_NN1 and_CC adds_VVZ some_DD specific_JJ information_NN1 on_II the_AT '_GE money_NN1 '_GE element_NN1 introduced_VVN by_II speaker_NN1 K._NP1 Speaker_NN1 K_ZZ221 's_ZZ222 next_MD contribution_NN1 exhibits_VVZ a_AT1 series_NN of_IO complex_JJ ties_NN2 with_IW the_AT existing_JJ topic_NN1 framework_NN1 ._. 
Speaker_NN1 J_ZZ1 ,_, in_II his_APPGE preceding_JJ contribution_NN1 ,_, has_VHZ talked_VVN about_II the_AT money_NN1 received_VVN for_IF his_APPGE work_NN1 at_II a_AT1 particular_JJ point_NN1 in_II the_AT past_NN1 ._. 
Speaker_NN1 K_ZZ221 's_ZZ222 contribution_NN1 '_GE picks_VVZ up_RP '_GE the_AT past_JJ time_NNT1 element_NN1 ,_, moving_VVG closer_RRR to_II speaker_NN1 J_ZZ221 's_ZZ222 time_VV0 while_CS maintaining_VVG the_AT personal_JJ reference_NN1 in_II my_APPGE father_NN1 ,_, who_PNQS also_RR did_VDD work_VVI (_( stonemason_NN1 )_) comparable_JJ to_TO J_ZZ221 's_ZZ222 (_( bricklayer_NN1 )_) and_CC received_VVD money_NN1 for_IF this_DD1 work_NN1 ._. 
Putting_VVG her_APPGE contribution_NN1 even_RR closer_RRR to_II J_ZZ1 's_VBZ preceding_JJ remarks_NN2 ,_, K_ZZ1 makes_VVZ her_APPGE comments_NN2 about_II her_APPGE father_NN1 relate_VV0 to_II his_58 '_GE starting_NN1 '_GE work_NN1 and_CC so_RG comparable_JJ to_TO J_ZZ221 's_ZZ222 started_VVD and_CC when_CS my_APPGE time_NNT1 was_VBDZ out_RP ._. 
With_IW these_DD2 complex_JJ connections_NN2 made_VVD ,_, speaker_NN1 K_ZZ1 adds_VVZ some_DD new_JJ elements_NN2 to_II the_AT conversation_NN1 (_( extra_JJ pay_NN1 for_IF being_VBG left-handed_RR )_) ._. 
We_PPIS2 have_VH0 tried_VVN to_TO list_VVI the_AT connections_NN2 existing_VVG across_II contributions_NN2 in_II this_DD1 discourse_NN1 fragment_NN1 to_TO emphasise_VVI the_AT ways_NN2 speakers_NN2 make_VV0 what_DDQ they_PPHS2 're_VBR talking_VVG about_II fit_JJ into_II a_AT1 framework_NN1 which_DDQ represents_VVZ what_DDQ we_PPIS2 (_( as_CSA discourse_NN1 participants_NN2 )_) are_VBR talking_VVG about_II in_II conversational_JJ discourse_NN1 ._. 
For_IF the_AT discourse_NN1 analyst_NN1 ,_, as_CSA an_AT1 overhearer_NN1 ,_, those_DD2 connections_NN2 can_VM signal_VVI the_AT coherence_NN1 relations_NN2 which_DDQ make_VV0 each_DD1 contribution_NN1 relevant_JJ to_II the_AT discourse_NN1 as_II a_AT1 whole_NN1 ._. 
Identifying_VVG the_AT elements_NN2 in_II the_AT topic_NN1 framework_NN1 at_II any_DD point_NN1 in_II the_AT discourse_NN1 allows_VVZ the_AT analyst_NN1 to_TO make_VVI claims_NN2 about_II what_DDQ is_VBZ involved_JJ in_II '_GE speaking_VVG topically_RR '_GE ._. 
It_PPH1 also_RR enables_VVZ him_PPHO1 to_TO produce_VVI a_AT1 version_NN1 of_IO '_" what_DDQ is_VBZ being_VBG talked_VVN about_II '_GE ,_, i.e._REX the_AT topic_NN1 of_IO conversation_NN1 ,_, which_DDQ is_VBZ much_RR more_RGR comprehensive_JJ ,_, and_CC certainly_RR of_IO greater_JJR analytic_JJ interest_NN1 ,_, than_CSN the_AT single_JJ word-or-phrase-type_JJ title_NN1 which_DDQ is_VBZ often_RR used_VVN in_II a_AT1 fairly_RR trivial_JJ way_NN1 to_TO characterise_VVI '_GE topic_NN1 '_GE in_II the_AT study_NN1 of_IO conversation_NN1 ._. 
Speaker_NN1 's_GE topic_NN1 So_RG far_RR we_PPIS2 have_VH0 considered_VVN the_AT notion_NN1 of_IO '_GE topic_NN1 '_GE in_II discourse_NN1 in_II31 terms_II32 of_II33 what_DDQ the_AT participants_NN2 share_VV0 ._. 
The_AT '_GE topic_NN1 framework_NN1 '_GE ,_, as_CSA an_AT1 analytic_JJ device_NN1 ,_, is_VBZ essentially_RR a_AT1 means_NN of_IO characterising_VVG the_AT area_NN1 of_IO overlap_NN1 in_II contributions_NN2 to_II a_AT1 discourse_NN1 ._. 
By_II concentrating_VVG on_II the_AT way_NN1 conversational_JJ contributions_NN2 overlap_VV0 ,_, however_RR ,_, we_PPIS2 may_VM neglect_VVI aspects_NN2 of_IO conversational_JJ discourse_NN1 associated_VVN with_IW different_JJ speakers_NN2 having_VHG different_JJ personal_JJ '_GE topics_NN2 '_GE ._. 
So_RG far_RR ,_, we_PPIS2 have_VH0 been_VBN concentrating_VVG on_II describing_VVG the_AT '_GE conversational_JJ topic_NN1 '_GE ,_, but_CCB neglecting_VVG the_AT notion_NN1 of_IO speaker_NN1 's_GE topic_NN1 ._. 
As_CSA we_PPIS2 have_VH0 already_RR pointed_VVN out_RP ,_, the_AT analyst_NN1 typically_RR treats_VVZ conversational_JJ data_NN as_CSA something_PN1 complete_JJ ,_, as_CSA a_AT1 static_JJ product_NN1 of_IO some_DD recorded_JJ interaction_NN1 ._. 
In_II doing_VDG so_RR ,_, he_PPHS1 may_VM lose_VVI sight_NN1 of_IO the_AT fact_NN1 that_CST conversational_JJ discourse_NN1 is_VBZ dynamic_JJ ,_, and_CC that_CST his_APPGE data_NN represents_VVZ a_AT1 process_NN1 ._. 
If_CS we_PPIS2 can_VM treat_VVI any_DD piece_NN1 of_IO conversational_JJ data_NN as_II a_AT1 process_NN1 in_II which_DDQ two_MC or_CC more_DAR participants_NN2 speak_VV0 within_II the_AT topic_NN1 framework_NN1 ,_, we_PPIS2 should_VM also_RR find_VVI in_II their_APPGE contributions_NN2 elements_NN2 which_DDQ characterise_VV0 their_APPGE own_DA personal_JJ '_GE speaker_NN1 's_GE topics_NN2 '_GE ._. 
We_PPIS2 shall_VM look_VVI at_II a_AT1 fragment_NN1 of_IO spoken_JJ discourse_NN1 ,_, not_XX in_II31 terms_II32 of_II33 how_RRQ we_PPIS2 would_VM characterise_VVI the_AT participants_NN2 '_GE shared_JJ information_NN1 ,_, but_CCB in_II31 terms_II32 of_II33 a_AT1 process_NN1 in_II which_DDQ each_DD1 participant_NN1 expresses_VVZ a_AT1 personal_JJ topic_NN1 within_II the_AT general_JJ topic_NN1 framework_NN1 of_IO the_AT conversation_NN1 as_II a_AT1 whole_NN1 ._. 
Prior_II21 to_II22 extract_NN1 (_( 11_MC )_) ,_, the_AT participants_NN2 ,_, L_ZZ1 (_( female_NN1 ,_, 20+_FO ,_, unmarried_JJ ,_, Edinburgh-resident_NN1 ,_, and_CC M_ZZ1 (_( female_NN1 ,_, 30+_FO ,_, married_JJ with_IW young_JJ children_NN2 ,_, Edinburgh-resident_NN1 )_) ,_, have_VH0 been_VBN talking_VVG about_II recent_JJ improvements_NN2 to_II old_JJ buildings_NN2 in_II different_JJ areas_NN2 in_II Edinburgh._NP1 (_( 11_MC )_) L_ZZ1 :_: I_PPIS1 quite_RR like_VV0 the_AT way_NN1 they_PPHS2 've_VH0 done_VDN the_AT Mile_NNU1 though_CS +_FO I_ZZ1 think_VV0 it_PPH1 's_VBZ quite_RG M_MC :_: yes&lsqb;AhA&rsqb;_VV0 yes_UH L_NP1 :_: the_AT bottom_NN1 of_IO it_PPH1 anyway_RR M_ZZ1 :_: it_PPH1 is_VBZ it_PPH1 is_VBZ quite_RG good_JJ they_PPHS2 've_VH0 certainly_RR kept_VVN within_II the_AT +_FO em_FU +_FO +_FO preserved_VVN it_PPH1 reasonably_RR well_RR or_CC conserved_VVD it_PPH1 but_CCB we_PPIS2 were_VBDR up_RP in_II Aberdeen_NP1 this_DD1 year_NNT1 for_IF a_AT1 holiday_NN1 and_CC we_PPIS2 were_VBDR staying_VVG right_RR within_II the_AT University_NN1 complex_NN1 there_RL in_II Old_JJ Aberdeen+_FO and_CC +_FO oh_UH some_DD of_IO the_AT buildings_NN2 there_EX are_VBR beautiful_JJ really_RR they_PPHS2 really_RR are_VBR nice_JJ +_FO but_CCB er_FU I_PPIS1 was_VBDZ quite_RR impressed_VVN with_IW it_PPH1 it_PPH1 's_VBZ the_AT first_MD holiday_NN1 we_PPIS2 've_VH0 had_VHN up_II there_RL +_FO L_ZZ1 :_: I_PPIS1 was_VBDZ noticing_VVG I_PPIS1 was_VBDZ down_RP by_II Queen_NNB Street_NP1 or+the_FO bottom_NN1 of_IO Hanover_NP1 Street_NNL1 or_CC somewhere_RL +_FO and_CC they_PPHS2 've_VH0 just_RR cleaned_VVN up_RP some_DD of_IO the_AT buildings_NN2 down_RP there_RL +_FO and_CC what_DDQ a_AT1 difference_NN1 it_PPH1 makes_VVZ +_FO M_ZZ1 :_: yes_UH I_PPIS1 know_VV0 because_CS there_EX are_VBR some_DD beautiful_JJ buildings_NN2 L_ZZ1 :_: oh_UH it_PPH1 was_VBDZ really_RR nice_JJ Extract_NN1 (_( 11_MC )_) is_VBZ representative_JJ of_IO a_AT1 common_JJ conversational_JJ situation_NN1 in_II which_DDQ each_DD1 of_IO the_AT participants_NN2 give_VV0 examples_NN2 from_II their_APPGE personal_JJ experience_NN1 to_TO illustrate_VVI some_DD general_JJ point_NN1 ._. 
The_AT general_JJ point_NN1 in_II this_DD1 case_NN1 is_VBZ something_PN1 like_II '_GE the_AT effect_NN1 of_IO restoring_VVG old_JJ buildings_NN2 '_GE which_DDQ is_VBZ already_RR part_NN1 of_IO the_AT topic_NN1 framework_NN1 established_VVN by_II the_AT preceding_JJ discourse_NN1 ._. 
Notice_VV0 that_DD1 speaker_NN1 M_MC 's_GE second_MD contribution_NN1 in_II this_DD1 extract_NN1 is_VBZ not_XX just_RR 'about_NN1 '_GE that_DD1 general_JJ point_NN1 ._. 
She_PPHS1 is_VBZ also_RR talking_VVG about_II her_APPGE recent_JJ holiday_NN1 in_II Aberdeen_NP1 ,_, for_REX21 example_REX22 ._. 
We_PPIS2 could_VM describe_VVI this_58 '_GE holiday_NN1 in_II Aberdeen_NP1 '_GE element_NN1 as_CSA ,_, at_II this_DD1 point_NN1 ,_, a_AT1 part_NN1 of_IO speaker_NN1 M_MC 's_VBZ personal_JJ topic_NN1 which_DDQ could_VM become_VVI ,_, in_II the_AT developing_JJ conversation_NN1 ,_, a_AT1 shared_JJ topic_NN1 area_NN1 for_IF both_DB2 speakers_NN2 ._. 
Speaker_NN1 L_ZZ1 could_VM have_VHI followed_VVN on_RP ,_, with_IW a_AT1 question_NN1 ,_, for_REX21 example_REX22 ,_, about_II the_AT holiday_NN1 ,_, Aberdeen_NP1 ,_, or_CC even_RR with_IW some_DD personal_JJ observations_NN2 on_II the_AT buildings_NN2 in_II Old_JJ Aberdeen_NP1 or_CC the_AT University_NN1 ._. 
Speaker_NN1 L_ZZ1 ,_, however_RR ,_, does_VDZ not_XX '_GE pick_VV0 up_RP '_" any_DD elements_NN2 from_II speaker_NN1 M_MC 's_VBZ personal_JJ topic_NN1 ,_, but_CCB continues_VVZ on_II her_APPGE own_DA personal_JJ topic_NN1 area_NN1 (_( i.e._REX Edinburgh_NP1 's_GE old_JJ buildings_NN2 after_II restoration_NN1 )_) ._. 
When_CS participant_NN1 M_ZZ1 speaks_VVZ again_RT near_II the_AT end_NN1 ,_, she_PPHS1 does_VDZ not_XX return_VVI to_II her_PPHO1 '_GE holiday_NN1 '_GE or_CC '_GE Old_JJ Aberdeen_NP1 '_GE ,_, but_CCB makes_VVZ her_APPGE contribution_NN1 relate_VVI closely_RR to_II L_ZZ1 's_VBZ immediately_RR preceding_JJ remarks_NN2 ._. 
There_EX are_VBR two_MC points_NN2 worth_II noting_VVG about_II this_DD1 fragment_NN1 of_IO conversational_JJ discourse_NN1 ._. 
First_MD ,_, it_PPH1 is_VBZ a_AT1 feature_NN1 of_IO a_AT1 lot_NN1 of_IO conversation_NN1 that_DD1 '_VBZ topics_NN2 '_GE are_VBR not_XX fixed_VVN beforehand_RR ,_, but_CCB are_VBR negotiated_VVN in_II the_AT process_NN1 of_IO conversing_VVG ._. 
Throughout_II a_AT1 conversation_NN1 ,_, the_AT next_MD '_GE topic_NN1 '_GE of_IO conversation_NN1 is_VBZ developing_VVG ._. 
Each_DD1 speaker_NN1 contributes_VVZ to_II the_AT conversation_NN1 in_II31 terms_II32 of_II33 both_DB2 the_AT existing_JJ topic_NN1 framework_NN1 and_CC his_APPGE or_CC her_APPGE personal_JJ topic_NN1 ._. 
It_PPH1 is_VBZ clear_JJ from_II extract_NN1 (_( 11_MC )_) that_CST some_DD elements_NN2 in_II a_AT1 speaker_NN1 's_GE personal_JJ topic_NN1 do_VD0 not_XX become_VVI salient_JJ elements_NN2 in_II the_AT conversation_NN1 if_CS neither_RR the_AT other_JJ participant_NN1 nor_CC the_AT speaker_NN1 herself_PPX1 mention_VV0 them_PPHO2 again_RT ._. 
To_TO use_VVI the_AT '_GE negotiation_NN1 '_GE metaphor_NN1 ,_, we_PPIS2 can_VM say_VVI that_DD1 speaker_NN1 M_ZZ1 offers_VVZ elements_NN2 in_II her_APPGE personal_JJ topic_NN1 (_( in_II her_APPGE second_MD contribution_NN1 )_) as_CSA possible_JJ elements_NN2 to_TO be_VBI included_VVN in_II the_AT conversational_JJ business_NN1 ,_, but_CCB speaker_NN1 L_ZZ1 does_VDZ not_XX take_VVI up_RP the_AT offer_NN1 ._. 
A_AT1 second_MD point_NN1 to_TO be_VBI noted_VVN in_II this_DD1 ,_, and_CC in_II a_AT1 large_JJ number_NN1 of_IO other_JJ conversational_JJ fragments_NN2 ,_, is_VBZ that_DD1 personal_JJ topics_NN2 are_VBR frequently_RR introduced_VVN through_II first_MD person_NN1 reference_NN1 in_II one_MC1 form_NN1 or_CC another_DD1 ._. 
Although_CS the_AT points_NN2 made_VVN in_II extract_NN1 (_( 11_MC )_) could_VM have_VHI been_VBN expressed_VVN objectively_RR as_CSA statements_NN2 that_CST certain_JJ buildings_NN2 in_II certain_JJ locations_NN2 are_VBR more_RGR beautiful_JJ since_CS restoration_NN1 ,_, both_DB2 speakers_NN2 relate_VV0 such_DA statements_NN2 to_II personal_JJ experience_NN1 ._. 
It_PPH1 is_VBZ as_CS21 if_CS22 speakers_NN2 feel_VV0 obliged_VVN to_TO offer_VVI some_DD personal_JJ warrant_NN1 for_IF the_AT statements_NN2 they_PPHS2 will_VM make_VVI about_II the_AT world_NN1 ._. 
A_AT1 statement_NN1 that_CST the_AT buildings_NN2 in_II Old_JJ Aberdeen_NP1 are_VBR beautiful_JJ is_VBZ embedded_VVN within_II an_AT1 assertion_NN1 that_CST the_AT speaker_NN1 was_VBDZ recently_RR in_II Old_JJ Aberdeen_NP1 ,_, and_CC stayed_VVD there_RL for_IF a_AT1 period_NN1 ,_, and_CC so_RR she_PPHS1 has_VHZ a_AT1 warrant_NN1 for_IF making_VVG the_AT statement_NN1 ._. 
If_CS we_PPIS2 reconsider_VV0 the_AT earlier_JJR extract_NN1 (_( 5_MC )_) as_CSA one_MC1 participant_NN1 wanting_VVG to_TO know_VVI the_AT meaning_NN1 of_IO an_AT1 expression_NN1 and_CC the_AT other_JJ offering_NN1 a_AT1 possible_JJ explanation_NN1 ,_, we_PPIS2 can_VM see_VVI that_CST the_AT explanation_NN1 is_VBZ offered_VVN in_II personal_JJ terms_NN2 (_( when_CS we_PPIS2 were_VBDR young_JJ and_CC we_PPIS2 called_VVD it_PPH1 '_VBZ the_AT taw_NN1 '_GE )_) based_VVN on_II the_AT speaker_NN1 's_GE personal_JJ experience_NN1 ._. 
It_PPH1 may_VM be_VBI that_CST this_DD1 explanation_NN1 is_VBZ not_XX an_AT1 acceptable_JJ answer_NN1 to_II the_AT question_NN1 ,_, but_CCB it_PPH1 is_VBZ presented_VVN by_II the_AT speaker_NN1 in_II a_AT1 form_NN1 which_DDQ conveys_58 '_GE what_DDQ I_PPIS1 think_VV0 we_PPIS2 're_VBR talking_VVG about_II '_GE in_II this_DD1 part_NN1 of_IO the_AT conversation_NN1 ._. 
Characterising_VVG the_AT individual_JJ speaker_NN1 's_GE topic_NN1 as_58 '_GE what_DDQ I_PPIS1 think_VV0 we_PPIS2 're_VBR talking_VVG about_II '_" incorporates_VVZ both_RR that_DD1 element_NN1 which_DDQ the_AT conversational_JJ analyst_NN1 tends_VVZ to_TO abstract_VVI as_II the_AT '_GE topic_NN1 of_IO conversation_NN1 '_GE for_IF the_AT participants_NN2 (_( 'What_VV0 we_PPIS2 're_VBR talking_VVG about'_NN1 )_) and_CC the_AT individual_JJ speaker_NN1 's_GE version_NN1 (_( 'I_UH think'_NN1 )_) ,_, as_CSA he/she_PPHS1 makes_VVZ a_AT1 conversational_JJ contribution_NN1 ._. 
That_DD1 speakers_NN2 do_VD0 introduce_VVI what_DDQ they_PPHS2 want_VV0 to_TO say_VVI via_II some_DD form_NN1 of_IO personal_JJ reference_NN1 has_VHZ a_AT1 noticeable_JJ effect_NN1 on_II the_AT structure_NN1 of_IO contributions_NN2 in_II conversational_JJ discourse_NN1 ._. 
We_PPIS2 shall_VM return_VVI to_II this_DD1 point_NN1 in_II the_AT discussion_NN1 of_IO further_JJR details_NN2 of_IO discourse_NN1 structure_NN1 in_II Chapter_NN1 4_MC ._. 
From_II what_DDQ we_PPIS2 have_VH0 proposed_VVN as_CSA speakers_NN2 '_GE topics_NN2 in_II conversational_JJ discourse_NN1 ,_, it_PPH1 must_VM occasionally_RR happen_VVI that_CST there_EX are_VBR at_RR21 least_RR22 two_MC versions_NN2 of_IO '_" What_DDQ I_PPIS1 think_VV0 we_PPIS2 're_VBR talking_VVG about_II '_" which_DDQ are_VBR potentially_RR incompatible_JJ ._. 
It_PPH1 is_VBZ a_AT1 noticeable_JJ feature_NN1 of_IO co-operative_JJ conversational_JJ discourse_NN1 ,_, however_RR ,_, that_CST this_DD1 potential_JJ incompatibility_NN1 rarely_RR leads_VVZ to_TO conflict_VVI over_II the_AT topic_NN1 of_IO conversation_NN1 ._. 
What_DDQ typically_RR happens_VVZ is_VBZ that_CST ,_, in_II the_AT negotiation_NN1 process_NN1 ,_, one_MC1 speaker_NN1 realises_VVZ that_CST his_APPGE version_NN1 is_VBZ incompatible_JJ with_IW what_DDQ the_AT other_NN1 appears_VVZ to_TO be_VBI talking_VVG about_II and_CC makes_VVZ his_APPGE contributions_NN2 compatible_JJ with_IW '_" what_DDQ I_PPIS1 think_VV0 you_PPY (_( not_XX we_PPIS2 )_) are_VBR talking_VVG about_II '_GE ._. 
We_PPIS2 can_VM illustrate_VVI this_DD1 process_NN1 in_II two_MC conversational_JJ fragments_NN2 and_CC note_VV0 two_MC different_JJ strategies_NN2 used_VMK to_TO avoid_VVI conflict_NN1 in_II the_AT '_GE negotiations_NN2 '_GE ._. 
In_II the_AT first_MD extract_NN1 ,_, (_( 12_MC )_) ,_, one_MC1 piece_NN1 of_IO continuous_JJ conversational_JJ discourse_NN1 has_VHZ been_VBN divided_VVN up_RP into_II chunks_NN2 ._. 
Immediately_RR before_CS this_DD1 extract_NN1 ,_, speaker_NN1 B_ZZ1 (_( female_NN1 ,_, 50+_FO ,_, aunt_NN1 of_IO speaker_NN1 A_ZZ1 )_) has_VHZ been_VBN describing_VVG to_II speaker_NN1 A_ZZ1 (_( female_NN1 ,_, 20+_FO )_) the_AT first_MD type_NN1 of_IO radio_NN1 she_PPHS1 had_VHD ,_, forty_MC years_NNT2 before._NNU (_( 12_MC )_) A_ZZ1 :_: but_CCB you_PPY 'd_VM have_VHI telephones_NN2 around_II +_FO B_ZZ1 :_: mm_UH oh_UH yes_UH oh_UH aye_UH oh_UH aye_UH I_PPIS1 've_VH0 had_VHN the_AT telephone_NN1 since_II nineteen_MC thirty_MC eight_MC +_FO A_ZZ1 :_: hmm_UH B_NP1 :_: oh_UH they_PPHS2 were_VBDR on_II a_AT1 long_JJ while_NNT1 I_PPIS1 think_VV0 before_II that_DD1 +_FO Speaker_NN1 B_ZZ1 had_VHD been_VBN talking_VVG about_II the_AT radio_NN1 she_PPHS1 had_VHD in_II the_AT 1930s_MC2 and_CC speaker_NN1 A_ZZ221 's_ZZ222 first_MD line_NN1 here_RL seems_VVZ to_TO continue_VVI within_II the_AT temporal_JJ ,_, locational_JJ and_CC personal_JJ indices_NN2 of_IO the_AT existing_JJ topic_NN1 framework_NN1 while_CS introducing_VVG telephones_NN2 ._. 
Speaker_NN1 B_ZZ1 treats_VVZ this_DD1 contribution_NN1 as_CSA requiring_VVG an_AT1 answer_NN1 ,_, following_VVG a_AT1 pattern_NN1 described_VVN by_II Labov_NP1 in_II the_AT rule_NN1 :_: '_GE If_CS (_( speaker_NN1 )_) A_ZZ1 makes_VVZ a_AT1 statement_NN1 about_II a_AT1 (_( speaker_NN1 )_) B-event_NN1 ,_, it_PPH1 is_VBZ heard_VVN as_II a_AT1 request_NN1 for_IF confirmation_NN1 '_GE (_( 1972b_FO :_: 254_MC )_) ._. 
Speaker_NN1 B_ZZ1 expands_VVZ on_II her_APPGE answer_NN1 ,_, in_II personal_JJ terms_NN2 ,_, regarding_VVG the_AT telephone_NN1 ._. 
Speaker_NN1 A_ZZ1 offers_VVZ no_AT contribution_NN1 and_CC speaker_NN1 B_ZZ1 adds_VVZ some_DD additional_JJ information_NN1 about_II telephones_NN2 ._. 
We_PPIS2 might_VM characterise_VVI speaker_NN1 B_ZZ221 's_ZZ222 view_VV0 of_IO '_" what_DDQ I_PPIS1 think_VV0 we_PPIS2 're_VBR talking_VVG about_II now_RT '_GE as_CSA something_PN1 involving_VVG herself_PPX1 ,_, the_AT 1930s_MC2 ,_, and_CC the_AT existence_NN1 of_IO telephones_NN2 (_( as_II31 well_II32 as_II33 radios_NN2 )_) at_II that_DD1 time_NNT1 ._. 
The_AT conversation_NN1 continues_VVZ :_: A_ZZ1 :_: 'cause_CS there_EX was_VBDZ a_AT1 man_NN1 in_II my_APPGE father_NN1 's_VBZ in_II the_AT Scouts_NN2 +_FO B_ZZ1 :_: oh_UH yes_UH he_PPHS1 was_VBDZ is_VBZ he_PPHS1 still_RR A_ZZ1 :_: he_PPHS1 's_VBZ a_AT1 county_NN1 commissioner_NN1 now_RT B_ZZ1 :_: oh_UH is_VBZ he_PPHS1 +ah_FO ha+_FO Speaker_NN1 A_ZZ1 appears_VVZ to_TO be_VBI offering_VVG some_DD new_JJ elements_NN2 as_II part_NN1 of_IO the_AT conversational_JJ topic_NN1 ,_, again_RT deriving_VVG from_II some_DD personal_JJ reference_NN1 (_( as_CSA in_II my_APPGE father_NN1 )_) which_DDQ speaker_NN1 B_ZZ1 appears_VVZ to_TO accept_VVI ._. 
That_REX21 is_REX22 ,_, speaker_NN1 B_ZZ1 does_VDZ not_XX insist_VVI on_II mentioning_VVG telephones_NN2 ,_, but_CCB moves_VVZ on_II21 to_II22 this_DD1 new_JJ area_NN1 ._. 
Speaker_NN1 B_ZZ221 's_ZZ222 view_VV0 of_IO '_" what_DDQ I_PPIS1 think_VV0 we_PPIS2 're_VBR talking_VVG about_II now_RT '_GE must_VM now_RT involve_VVI speaker_NN1 A_ZZ1 ,_, A_ZZ221 's_ZZ222 father_VV0 ,_, the_AT Scouts_NN2 and_CC a_AT1 man_NN1 (_( who_PNQS may_VM have_VHI something_PN1 to_TO do_VDI with_IW telephones_NN2 )_) ._. 
We_PPIS2 might_VM expect_VVI speaker_NN1 B_ZZ1 to_TO be_VBI a_RR21 little_RR22 confused_VVN about_II how_RRQ these_DD2 elements_NN2 relate_VV0 to_II the_AT preceding_JJ conversation_NN1 ._. 
Speaker_NN1 A_ZZ1 continues_VVZ ,_, as_CSA follows_VVZ :_: A_ZZ1 :_: and_CC eh_UH one_MC1 of_IO his_PPGE oldest+_FO scoutmasters_NN2 wa-ha-was_VBDZ reaching_VVG his_APPGE hundredth_MD birthday_NN1 +_FO B_ZZ1 :_: is_VBZ that_DD1 so_RR +_FO We_PPIS2 suspect_VV0 that_CST ,_, by_II this_DD1 point_NN1 ,_, although_CS speaker_NN1 B_ZZ1 can_VM identify_VVI '_" what_DDQ 's_VBZ being_VBG talked_VVN about_II '_GE ,_, she_PPHS1 can_VM play_VVI no_AT part_NN1 in_II negotiating_VVG the_AT topic_NN1 ,_, because_CS she_PPHS1 may_VM not_XX be_VBI able_JK to_TO see_VVI why_RRQ this_DD1 individual_JJ entity_NN1 is_VBZ being_VBG talked_VVN about_II ._. 
The_AT contributions_NN2 of_IO speaker_NN1 B_ZZ1 cease_VV0 to_TO be_VBI attempts_NN2 to_TO add_VVI anything_PN1 to_II the_AT conversational_JJ topic_NN1 ._. 
Speaker_NN1 B_ZZ221 's_ZZ222 view_VV0 of_IO the_AT conversation_NN1 has_VHZ consequently_RR become_VV0 one_PN1 in_II which_DDQ she_PPHS1 is_VBZ no_RR21 longer_RR22 expressing_VVG a_AT1 personal_JJ topic_NN1 ,_, but_CCB is_VBZ waiting_VVG to_TO discover_VVI '_" what_DDQ I_PPIS1 think_VV0 you_PPY (_( not_XX we_PPIS2 )_) are_VBR talking_VVG about_II '_GE ._. 
Throughout_II the_AT rest_NN1 of_IO this_DD1 fragment_NN1 ,_, speaker_NN1 B_ZZ1 simply_RR makes_NN2% '_GE interested_JJ '_GE noises_NN2 as_CSA speaker_NN1 A_ZZ1 gradually_RR gets_VVZ to_II the_AT point_NN1 ._. 
A_ZZ1 :_: so_RR father_NN1 was_VBDZ making_VVG up_RP a_AT1 big_JJ +_FO sort_NN1 of_IO remembrance_NN1 book_NN1 B_ZZ1 :_: aha_UH A_ZZ1 :_: to_TO give_VVI him_PPHO1 and_CC he_PPHS1 was_VBDZ writing_VVG just_RR at_II the_AT beginning_NN1 he_PPHS1 was_VBDZ writing_VVG the_AT whole_NN1 for_IF each_DD1 year_NNT1 of_IO his_APPGE life_NN1 he_PPHS1 wrote_VVD something_PN1 in_II that_DD1 had_VHD had_VHN been_VBN invented_VVN or_CC +_FO B_ZZ1 :_: oh_UH yes_UH A_ZZ1 :_: ah_UH a_AT1 book_NN1 that_CST had_VHD been_VBN written_VVN or_CC a_AT1 piece_NN1 of_IO music_NN1 that_CST had_VHD been_VBN written_VVN or_CC a_AT1 painting_NN1 or_CC a_AT1 B_ZZ1 :_: very_RG interesting_JJ yes_UH A_ZZ1 :_: or_CC whatever_DDQV you_PPY know_VV0 and_CC +_FO within_II his_APPGE lifetime_NNT1 the_AT telephone_NN1 had_VHD been_VBN invented_VVN +_FO B_ZZ1 :_: had_VHD it_PPH1 +_FO really_RR +_FO fancy_VV0 +_FO In_II this_DD1 extract_NN1 as_II a_AT1 whole_NN1 ,_, we_PPIS2 can_VM trace_VVI speaker_NN1 B_ZZ221 's_ZZ222 attempt_VV0 to_TO contribute_VVI to_II what_DDQ she_PPHS1 thinks_VVZ they_PPHS2 're_VBR about_RP ,_, by_II first_MD offering_VVG some_DD remarks_NN2 on_II telephones_NN2 and_CC then_RT on_II the_AT father_NN1 ,_, but_CCB gradually_RR reducing_VVG her_APPGE comments_NN2 to_II the_AT type_NN1 of_IO contentless_JJ noises_NN2 described_VVN by_II Duncan_NP1 (_( 1973_MC )_) as_CSA back_JJ channels_NN2 ._. 
Back_NN1 channel_NN1 behaviour_NN1 ,_, which_DDQ can_VM also_RR include_VVI nods_NN2 and_CC sentence_NN1 completions_NN2 is_VBZ used_VVN when_CS a_AT1 participant_NN1 wants_VVZ to_TO indicate_VVI to_II the_AT person_NN1 speaking_VVG that_CST he_PPHS1 should_VM continue_VVI ._. 
Speaker_NN1 B_ZZ1 stops_VVZ trying_VVG to_TO take_VVI turns_NN2 in_II the_AT negotiation_NN1 of_IO topic_NN1 and_CC waits_VVZ for_IF speaker_NN1 A_ZZ1 to_TO make_VVI it_PPH1 clear_JJ how_RRQ what_DDQ she_PPHS1 is_VBZ saying_VVG has_VHZ some_DD connection_NN1 to_II the_AT existing_JJ topic_NN1 framework_NN1 ._. 
Eventually_RR ,_, as_CSA we_PPIS2 can_VM see_VVI in_II A_ZZ1 's_VBZ final_JJ remarks_NN2 ,_, a_AT1 connection_NN1 is_VBZ made_VVN ._. 
There_EX is_VBZ evidence_NN1 in_II speaker_NN1 A_ZZ221 's_ZZ222 contributions_NN2 that_CST what_DDQ she_PPHS1 is_VBZ trying_VVG to_TO say_VVI is_VBZ not_XX very_RG well_RR organised_VVN before_CS she_PPHS1 starts_VVZ to_TO speak_VVI ._. 
There_EX are_VBR false_RR starts_VVZ ,_, hesitations_NN2 and_CC repetitions_NN2 ._. 
Everyday_JJ conversational_JJ discourse_NN1 is_VBZ ,_, not_XX infrequently_RR ,_, characterised_VVN by_II this_DD1 lack_NN1 of_IO preplanning_VVG ._. 
The_AT resulting_JJ structure_NN1 of_IO speaker_NN1 A_ZZ221 's_ZZ222 contributions_NN2 is_VBZ ,_, in_II fact_NN1 ,_, quite_RG common_JJ in_II discourse_NN1 and_CC will_VM be_VBI discussed_VVN in_II some_DD detail_NN1 later_RRR in_II31 terms_II32 of_II33 '_GE staging_NN1 '_GE (_( see_VV0 Chapter_NN1 4_MC )_) ._. 
Speaker_NN1 B_ZZ221 's_ZZ222 strategy_NN1 ,_, then_RT ,_, in_II a_AT1 situation_NN1 where_RRQ she_PPHS1 finds_VVZ that_CST she_PPHS1 is_VBZ unsure_JJ about_II what_DDQ she_PPHS1 thinks_VVZ they_PPHS2 're_VBR talking_VVG about_II ,_, is_VBZ to_TO stop_VVI talking_VVG ._. 
In_II the_AT following_JJ extract_NN1 (_( 13_MC )_) ,_, there_EX is_VBZ another_DD1 example_NN1 of_IO a_AT1 mismatch_NN1 between_II speakers_NN2 '_GE topics_NN2 ,_, brought_VVN about_RP by_II a_AT1 misunderstanding_NN1 of_IO the_AT intended_JJ meaning_NN1 of_IO a_AT1 particular_JJ word_NN1 ._. 
In_II the_AT immediately_RR preceding_JJ conversation_NN1 ,_, speaker_NN1 C_ZZ1 (_( female_NN1 ,_, 20+_FO ,_, American_JJ ,_, visiting_VVG Edinburgh_NP1 )_) has_VHZ been_VBN finding_VVG out_RP from_II speaker_NN1 D_ZZ1 (_( male_NN1 ,_, 40+_FO ,_, Edinburgh-resident_NN1 )_) where_RRQ there_EX are_VBR good_JJ places_NN2 to_TO go_VVI for_IF bicycle_NN1 rides_VVZ in_RP and_CC around_II Edinburgh._NP1 (_( 13_MC )_) C_ZZ1 :_: what_DDQ about_II going_VVG down_RP by_II the_AT the_AT Firth_NN1 of_IO Forth_NP1 D_NP1 :_: that_DD1 should_VM be_VBI fun_NN1 should_VM n't_XX it_PPH1 yes_UH you_PPY could_VM C_ZZ1 :_: is_VBZ it_PPH1 D_NP1 :_: yes_UH you_PPY can_VM cycle_NN1 all_DB you_PPY can_VM ride_VVI right_RR along_II the_AT edge_NN1 you_PPY know_VV0 +_FO without_IW falling_VVG in_II you_PPY can_VM ride_VVI right_RR along_II the_AT edge_NN1 eh_UH without_RR em_FU +_FO going_VVG keeping_VVG on_II the_AT main_JJ road_NN1 +_FO that_DD1 should_VM be_VBI great_JJ actually_RR +_FO you_PPY could_VM do_VDI that_DD1 +_FO C_ZZ1 :_: is_VBZ it_PPH1 very_RG rough_JJ down_RP there_RL though_CS D_ZZ1 :_: well_RR there_EX are_VBR no_AT cobbles_NN2 as_CS31 far_CS32 as_CS33 I_PPIS1 remember_VV0 have_VH0 you_PPY tried_VVN riding_VVG on_II the_AT cobbles_NN2 C_ZZ1 :_: yes_UH yes_UH D_NP1 :_: you_PPY must_VM have_VHI done_VDN C_ZZ1 :_: I_PPIS1 went_VVD down_RP to_II Muirhouse_NP1 D_NP1 :_: which_DDQ is_VBZ almost_RR all_DB cobbles_NN2 is_VBZ n't_XX it_PPH1 C_NP1 :_: it_PPH1 was_VBDZ rather_RG rough_JJ D_ZZ1 :_: hmm_UH C_NP1 :_: no_UH but_CCB I_PPIS1 was_VBDZ I_PPIS1 was_VBDZ thinking_VVG rather_RG more_RGR rough_JJ in_II31 terms_II32 of_II33 the_AT em_FU +_FO people_NN +_FO D_ZZ1 :_: oh_UH I_PPIS1 see_VV0 +_FO you_PPY well_RR I_PPIS1 do_VD0 n't_XX think_VVI so_RG +_FO I_PPIS1 do_VD0 n't_XX know_VVI +_FO I_ZZ1 I_ZZ1 eh_UH parts_NN2 of_IO it_PPH1 are_VBR quite_RG poor_JJ +_FO particularly_RR the_AT Pilton_NP1 area_NN1 +_FO Looking_VVG back_RP to_II speaker_NN1 C_ZZ221 's_ZZ222 third_MD question_NN1 ,_, we_PPIS2 can_VM propose_VVI two_MC versions_NN2 of_IO '_" what_DDQ I_PPIS1 think_VV0 we_PPIS2 're_VBR talking_VVG about_II '_GE ._. 
For_IF speaker_NN1 C_ZZ1 ,_, it_PPH1 involves_58 '_GE are_VBR the_AT people_NN rough_VV0 ?_? '_GE and_CC ,_, for_IF speaker_NN1 D_ZZ1 ,_, '_" are_VBR the_AT roads_NN2 rough_VV0 ?_? 
'_GE Unlike_JJ the_AT hearer_NN1 (_( B_ZZ1 )_) in_II extract_NN1 (_( 12_MC )_) ,_, however_RR ,_, speaker_NN1 C_ZZ1 appears_VVZ to_TO be_VBI able_JK to_TO recognise_VVI speaker_NN1 D_ZZ1 's_VBZ alternative_JJ topic_NN1 and_CC accepts_VVZ what_DDQ she_PPHS1 thinks_VVZ speaker_NN1 D_ZZ1 is_VBZ talking_VVG about_II as_58 '_GE what_DDQ we_PPIS2 're_VBR talking_VVG about_II '_GE ,_, for_IF a_AT1 few_DA2 turns_NN2 ._. 
When_CS speaker_NN1 D_ZZ1 stops_VVZ talking_VVG about_II cobbles_NN2 (_( i.e._REX rough_JJ for_IF cycling_VVG on_RP )_) ,_, speaker_NN1 C_ZZ1 can_VM attempt_VVI to_TO return_VVI to_II her_APPGE topic_NN1 (_( rough_VV0 in_II31 terms_II32 of_II33 the_AT em_FU +_FO people_NN )_) ._. 
Speaker_NN1 D_ZZ221 's_ZZ222 response_NN1 at_II the_AT end_NN1 of_IO this_DD1 fragment_NN1 is_VBZ ,_, in_II effect_NN1 ,_, an_AT1 answer_NN1 to_II the_AT question_NN1 which_DDQ speaker_NN1 C_ZZ1 originally_RR intended_VVD him_PPHO1 to_TO answer_VVI ._. 
We_PPIS2 might_VM think_VVI that_CST by_II the_AT end_NN1 of_IO this_DD1 fragment_NN1 there_EX is_VBZ once_RR21 again_RR22 a_AT1 single_JJ version_NN1 for_IF both_DB2 speakers_NN2 of_IO '_" what_DDQ I_PPIS1 think_VV0 we_PPIS2 're_VBR talking_VVG about_II '_GE ._. 
Indeed_RR ,_, most_RGT conversational_JJ analysis_NN1 is_VBZ undertaken_VVN with_IW this_DD1 single_NN1 '_GE topic_NN1 '_GE concept_NN1 as_II a_AT1 working_JJ assumption_NN1 ._. 
Yet_RR ,_, in_II extract_NN1 (_( 13_MC )_) we_PPIS2 can_VM only_RR reconstruct_VVI the_AT intended_JJ meaning_NN1 of_IO C_ZZ221 's_ZZ222 third_MD question_NN1 because_CS she_PPHS1 actually_RR explains_VVZ her_PPHO1 intended_VVD meaning_VVG later_RRR ._. 
If_CS speaker_NN1 D_ZZ1 had_VHD gone_VVN on_RP at_II some_DD length_NN1 about_II '_GE cobbles_NN2 '_GE or_CC rough_JJ roads_NN2 in_RR21 general_RR22 ,_, or_CC if_CS the_AT analysis_NN1 only_RR had_VHN part_NN1 of_IO this_DD1 fragment_NN1 ,_, up_RP to_II C_ZZ1 's_VBZ it_PPH1 was_VBDZ rather_RG rough_JJ ,_, then_RT we_PPIS2 might_VM have_VHI had_VHN no_AT evidence_NN1 of_IO a_AT1 divergence_NN1 in_II speakers_NN2 '_GE topics_NN2 within_II the_AT conversation_NN1 ._. 
Our_APPGE argument_NN1 for_IF the_AT importance_NN1 of_IO considering_VVG individual_JJ speaker_NN1 's_GE topics_NN2 in_II conversational_JJ discourse_NN1 would_VM consequently_RR be_VBI weaker_JJR ._. 
We_PPIS2 do_VD0 not_XX suggest_VVI that_DD1 discourse_NN1 analysts_NN2 should_VM spend_VVI their_APPGE time_NNT1 looking_VVG for_IF potential_JJ alternative_JJ meanings_NN2 in_II what_DDQ speakers_NN2 say_VV0 in_II a_AT1 conversation_NN1 ,_, but_CCB we_PPIS2 do_VD0 suggest_VVI that_CST the_AT analyst_NN1 should_VM not_XX simply_RR assume_VVI that_CST there_EX is_VBZ a_AT1 single_JJ ,_, static_NN1 '_GE topic_NN1 of_IO conversation_NN1 '_GE in_II any_DD conversational_JJ fragment_NN1 ._. 
If_CS there_EX is_VBZ an_AT1 entity_NN1 identifiable_JJ as_58 '_GE the_AT topic_NN1 of_IO conversation_NN1 '_GE ,_, the_AT analyst_NN1 should_VM consider_VVI what_DDQ evidence_NN1 from_II each_DD1 individual_JJ speaker_NN1 's_GE contributions_NN2 he_PPHS1 is_VBZ using_VVG to_TO make_VVI that_DD1 identification_NN1 ._. 
He_PPHS1 should_VM also_RR remain_VVI aware_JJ of_IO the_AT fact_NN1 that_CST conversation_NN1 is_VBZ a_AT1 process_NN1 and_CC that_CST each_DD1 contribution_NN1 should_VM be_VBI treated_VVN as_II part_NN1 of_IO the_AT negotiation_NN1 of_IO '_" what_DDQ is_VBZ being_VBG talked_VVN about_II '_GE ._. 
Above_II all_DB ,_, he_PPHS1 should_VM remember_VVI that_CST it_PPH1 is_VBZ speakers_NN2 ,_, and_CC not_XX conversations_NN2 or_CC discourses_NN2 ,_, that_CST have_VH0 '_GE topics_NN2 '_GE ._. 
Topic_NN1 boundary_NN1 markers_NN2 In_II our_APPGE discussion_NN1 of_IO '_GE topic_NN1 '_GE ,_, we_PPIS2 have_VH0 concentrated_VVN mainly_RR on_II considerations_NN2 of_IO '_GE content_NN1 '_GE and_CC neglected_VVD the_AT influence_NN1 of_IO '_GE form_NN1 '_GE ._. 
Yet_RR our_APPGE interpretation_NN1 of_IO what_DDQ a_AT1 speaker_NN1 is_VBZ talking_VVG about_II is_VBZ inevitably_RR based_VVN on_II how_RRQ he_PPHS1 structures_VVZ what_DDQ he_PPHS1 is_VBZ saying_VVG ._. 
We_PPIS2 shall_VM now_RT investigate_VVI some_DD formal_JJ aspects_NN2 of_IO topic-structure_NN1 in_II discourse_NN1 ._. 
In_II this_DD1 section_NN1 we_PPIS2 shall_VM look_VVI at_II the_AT formal_JJ devices_NN2 used_VMK to_TO mark_VVI the_AT boundaries_NN2 of_IO chunks_NN2 of_IO both_DB2 written_VVN and_CC spoken_JJ discourse_NN1 which_DDQ form_VV0 large_JJ units_NN2 of_IO some_DD kind_NN1 ,_, such_II21 as_II22 paragraphs_NN2 ._. 
Aspects_NN2 of_IO the_AT internal_JJ structuring_NN1 of_IO these_DD2 chunks_NN2 will_VM be_VBI discussed_VVN in_II Chapter_NN1 4_MC ._. 
It_PPH1 has_VHZ been_VBN suggested_VVN (_( e.g._REX by_II Schank_NP1 ,_, 1977_MC :_: 424_MC ;_; Maynard_NP1 ,_, 1980_MC )_) that_CST instead_II21 of_II22 undertaking_VVG the_AT difficult_JJ task_NN1 of_IO attempting_VVG to_TO define_VVI '_" what_DDQ a_AT1 topic_NN1 is_58 '_GE ,_, we_PPIS2 should_VM concentrate_VVI on_II describing_VVG what_DDQ we_PPIS2 recognise_VV0 as_II topic-shift_NN1 ._. 
That_REX21 is_REX22 ,_, between_II two_MC contiguous_JJ pieces_NN2 of_IO discourse_NN1 which_DDQ are_VBR intuitively_RR considered_VVN to_TO have_VHI two_MC different_JJ '_GE topics_NN2 '_GE ,_, there_EX should_VM be_VBI a_AT1 point_NN1 at_II which_DDQ the_AT shift_NN1 from_II one_MC1 topic_NN1 to_II the_AT next_MD is_VBZ marked_VVN ._. 
If_CS we_PPIS2 can_VM characterise_VVI this_DD1 marking_NN1 of_IO topic-shift_NN1 ,_, then_RT we_PPIS2 shall_VM have_VHI found_VVN a_AT1 structural_JJ basis_NN1 for_IF dividing_VVG up_RP stretches_NN2 of_IO discourse_NN1 into_II a_AT1 series_NN of_IO smaller_JJR units_NN2 ,_, each_DD1 on_II a_AT1 separate_JJ topic_NN1 ._. 
This_DD1 type_NN1 of_IO approach_NN1 to_II the_AT analysis_NN1 of_IO discourse_NN1 is_VBZ based_VVN on_II the_AT principle_NN1 that_CST ,_, if_CS we_PPIS2 can_VM identify_VVI the_AT boundaries_NN2 of_IO units_NN2 where_CS one_MC1 unit_NN1 ends_NN2 and_CC another_DD1 begins_VVZ then_RT we_PPIS2 need_VM not_XX have_VHI a_JJ21 priori_JJ22 specifications_NN2 for_IF the_AT content_NN1 of_IO such_DA units_NN2 ._. 
The_AT burden_NN1 of_IO analysis_NN1 is_VBZ consequently_RR transferred_VVN to_II identifying_VVG the_AT formal_JJ markers_NN2 of_IO topic-shift_NN1 in_II discourse_NN1 ._. 
Paragraphs_NN2 It_PPH1 might_VM seem_VVI that_CST identifying_VVG the_AT formal_JJ demarcation_NN1 of_IO chunks_NN2 of_IO written_JJ or_CC printed_JJ discourse_NN1 is_VBZ a_AT1 relatively_RR simple_JJ task_NN1 ._. 
After_II all_DB ,_, written_JJ discourse_NN1 is_VBZ divided_VVN into_II paragraphs_NN2 whose_DDQGE boundaries_NN2 are_VBR marked_VVN by_II indentations_NN2 ._. 
Topic-shifts_NN2 in_II written_JJ discourse_NN1 then_RT could_VM be_VBI identified_VVN with_IW the_AT beginning_NN1 of_IO each_DD1 new_JJ paragraph_NN1 ._. 
Unfortunately_RR ,_, it_PPH1 does_VDZ n't_XX seem_VVI to_TO be_VBI as_RG simple_JJ as_CSA that_DD1 ._. 
Those_DD2 who_PNQS use_VV0 the_AT term_NN1 '_GE paragraph_NN1 '_GE to_TO describe_VVI a_AT1 unit_NN1 in_II the_AT structural_JJ analysis_NN1 of_IO written_JJ discourse_NN1 go_VV0 to_II some_DD trouble_NN1 to_TO point_VVI out_RP that_CST they_PPHS2 are_VBR not_XX describing_VVG the_AT orthographic_JJ paragraph_NN1 ._. 
According_II21 to_II22 Longacre_NP1 (_( 1979_MC :_: 116_MC )_) ,_, the_AT orthographic_JJ paragraph_NN1 can_VM result_VVI from_II a_AT1 writer_NN1 's_GE stylistic_JJ concerns_NN2 ,_, '_GE partially_RR dictated_VVN by_II eye_NN1 appeal_NN1 '_GE ,_, or_CC from_II printing_VVG conventions_NN2 such_II21 as_II22 an_AT1 indentation_NN1 for_IF each_DD1 change_NN1 of_IO speaker_NN1 ._. 
Hinds_NP1 (_( 1977_MC :_: 83_MC )_) also_RR notes_VVZ that_CST the_AT journalistic_JJ paragraph_NN1 is_VBZ often_RR determined_VVN on_II the_AT basis_NN1 of_IO appearance_NN1 ._. 
He_PPHS1 has_VHZ a_AT1 worked_JJ example_NN1 in_II which_DDQ a_AT1 single_JJ structural_JJ paragraph_NN1 derives_VVZ from_II a_AT1 newspaper_NN1 article_NN1 containing_VVG five_MC orthographic_JJ paragraphs_NN2 ._. 
Thus_RR ,_, it_PPH1 may_VM be_VBI that_CST the_AT beginning_NN1 of_IO an_AT1 orthographic_JJ paragraph_NN1 indicates_VVZ a_AT1 point_NN1 of_IO topic-shift_NN1 ,_, but_CCB it_PPH1 need_VM not_XX do_VDI so_RR ._. 
Both_DB2 Longacre_NP1 (_( 1979_MC )_) and_CC Hinds_NP1 (_( 1977_MC )_) appeal_VV0 to_II languages_NN2 other_II21 than_II22 English_NN1 for_IF evidence_NN1 that_CST there_EX are_VBR formal_JJ linguistic_JJ markers_NN2 of_IO the_AT beginning_NN1 and_CC end_NN1 of_IO paragraphs_NN2 ._. 
What_DDQ is_VBZ immediately_RR noticeable_JJ in_II the_AT discussion_NN1 of_IO these_DD2 markers_NN2 is_VBZ that_CST they_PPHS2 are_VBR genre-specific_JJ ._. 
There_EX are_VBR ways_NN2 of_IO indicating_VVG the_AT beginning_NN1 of_IO a_AT1 new_JJ paragraph_NN1 in_II a_AT1 piece_NN1 of_IO narrative_NN1 ,_, for_REX21 example_REX22 ,_, which_DDQ are_VBR not_XX used_VVN in_II explanatory_JJ discourse_NN1 ._. 
This_DD1 general_JJ point_NN1 is_VBZ also_RR made_VVN by_II Grimes_NP1 (_( 1975_MC :_: 109_MC )_) ,_, who_PNQS describes_VVZ the_AT marking_NN1 of_IO paragraph_NN1 boundaries_NN2 as_CSA one_MC1 form_NN1 of_IO '_GE partitioning_NN1 '_GE in_II discourse_NN1 ._. 
The_AT principles_NN2 on_II which_DDQ partitioning_NN1 depends_VVZ are_VBR related_VVN to_TO change_VVI of_IO '_GE setting_NN1 '_GE (_( time_NNT1 or_CC place_NN1 )_) and_CC '_GE theme_NN1 '_GE (_( the_AT person_NN1 or_CC thing_NN1 talked_VVN about_II )_) ,_, in_II narrative_JJ discourse_NN1 ,_, at_RR21 least_RR22 ._. 
Interesting_JJ though_CS it_PPH1 may_VM be_VBI to_TO learn_VVI that_CST there_EX is_VBZ a_AT1 in_II Huichol_NP1 or_CC Shipibo_NP1 ,_, it_PPH1 becomes_VVZ decidedly_RR less_RGR interesting_JJ when_CS one_PN1 discovers_VVZ that_CST the_AT identification_NN1 of_IO the_AT significance_NN1 of_IO these_DD2 particles_NN2 depends_VVZ on_II a_AT1 prior_JJ identification_NN1 of_IO the_AT paragraph_NN1 as_II a_AT1 unit_NN1 in_II which_DDQ '_VBZ the_AT speaker_NN1 continues_VVZ talking_VVG about_II the_AT same_DA thing_NN1 '_GE (_( Grimes_NP1 ,_, 1975_MC :_: 103_MC )_) ._. 
Hinds_NP1 (_( 1977_MC )_) bases_VVZ his_APPGE paragraph_NN1 divisions_NN2 on_II a_AT1 similar_JJ principle_NN1 ,_, quoting_VVG Grimes_NP1 as_CSA support_NN1 ,_, and_CC emphasising_VVG the_AT significance_NN1 of_IO '_GE participant_NN1 orientation_NN1 '_GE that_REX21 is_REX22 ,_, the_AT unity_NN1 of_IO a_AT1 paragraph_NN1 derives_VVZ from_II its_APPGE being_VBG mainly_RR about_II a_AT1 single_JJ participant_NN1 ._. 
Longacre_NP1 (_( 1979_MC )_) claims_VVZ that_DD1 '_VBZ in_II narrative_JJ discourse_NN1 ,_, a_AT1 narrative_JJ paragraph_NN1 is_VBZ built_VVN around_II a_AT1 thematic_JJ participant_NN1 ,_, occasionally_RR a_AT1 small_JJ set_NN1 of_IO thematic_JJ participants_NN2 '_GE (_( Longacre_NP1 ,_, 1979_MC :_: 118_MC )_) ._. 
In_II other_JJ words_NN2 ,_, only_RR the_AT paragraph_NN1 structure_NN1 of_IO stretches_NN2 of_IO discourse_NN1 about_II individual_NN1 ,_, primarily_RR human_JJ ,_, characters_NN2 is_VBZ being_VBG discussed_VVN ._. 
In_II effect_NN1 ,_, this_DD1 limits_VVZ the_AT discussion_NN1 to_II narrative_JJ discourse_NN1 ,_, or_CC ,_, as_CSA in_II Hinds_NP1 (_( 1977_MC )_) ,_, a_AT1 description_NN1 or_CC an_AT1 obituary_NN1 of_IO a_AT1 particular_JJ individual_NN1 ._. 
It_PPH1 should_VM be_VBI obvious_JJ why_RRQ a_AT1 single_JJ structural_JJ or_CC '_GE semantic_JJ '_GE paragraph_NN1 in_II Hinds_NP1 '_GE (_( 1977_MC )_) analysis_NN1 can_VM extend_VVI over_RG five_MC orthographic_JJ paragraphs_NN2 in_II a_AT1 newspaper_NN1 ._. 
Each_DD1 of_IO these_DD2 orthographic_JJ paragraphs_NN2 is_58 '_GE about_II '_GE the_AT same_DA individual_NN1 ._. 
Yet_RR ,_, some_DD obituaries_NN2 extend_VV0 to_II twenty_MC or_CC more_RGR orthographic_JJ paragraphs_NN2 '_GE about_II '_GE the_AT one_MC1 person_NN1 ,_, and_CC whole_JJ chapters_NN2 of_IO novels_NN2 ,_, containing_VVG over_RG a_AT1 hundred_NNO lengthy_JJ orthographic_JJ paragraphs_NN2 ,_, may_VM be_VBI '_GE about_II '_GE the_AT same_DA individual_NN1 ._. 
Surely_RR such_DA extended_JJ stretches_NN2 of_IO written_JJ discourse_NN1 are_VBR not_XX single_JJ '_GE paragraphs_NN2 '_GE ?_? 
We_PPIS2 shall_VM consider_VVI a_AT1 stretch_NN1 of_IO written_JJ discourse_NN1 ,_, not_XX from_II a_AT1 source_NN1 such_II21 as_II22 a_AT1 Paez_NP1 (_( Colombia_NP1 )_) folk_NN tale_NN1 or_CC a_AT1 specially_RR constructed_VVN text_NN1 ,_, but_CCB from_II a_AT1 recent_JJ English_JJ novel_NN1 ._. 
In_II the_AT extract_NN1 reproduced_VVD below_RL (_( 14_MC )_) ,_, the_AT orthographic_JJ paragraph_NN1 boundaries_NN2 as_CSA they_PPHS2 appeared_VVD on_II the_AT printed_JJ page_NN1 have_VH0 been_VBN ignored_VVN ._. 
The_AT whole_JJ extract_NN1 has_VHZ two_MC principal_JJ participants_NN2 ,_, but_CCB is_VBZ quite_RG clearly_RR '_GE about_II '_GE only_RR one_MC1 of_IO them_PPHO2 ._. 
If_CS there_EX are_VBR points_NN2 of_IO '_GE topic-shift_NN1 '_GE in_II English_JJ written_JJ discourse_NN1 which_DDQ lead_VV0 writers_NN2 ,_, or_CC their_APPGE editors_NN2 ,_, to_TO begin_VVI new_JJ orthographic_JJ paragraphs_NN2 ,_, then_RT we_PPIS2 should_VM be_VBI able_JK to_TO identify_VVI likely_JJ points_NN2 where_RRQ the_AT writer_NN1 or_CC the_AT editor_NN1 marked_VVN the_AT division_NN1 of_IO this_58 '_GE text_NN1 '_GE into_II separate_JJ chunks._NNU (_( 14_MC )_) 1After_FO the_AT first_MD few_DA2 days_NNT2 ,_, when_RRQ I_PPIS1 come_VV0 into_II the_AT room_NN1 ,_, Birdie_NP1 is_VBZ down_RP on_II the_AT floor_NN1 of_IO the_AT cage_NN1 ,_, running_VVG back_RP and_CC forth_RR ,_, looking_VVG out_RP over_II the_AT barrier_NN1 that_CST holds_VVZ in_II the_AT gravel._NNU 2_MC I_PPIS1 think_VV0 she_PPHS1 's_VBZ glad_JJ to_TO see_VVI me_PPIO1 ,_, not_XX just_RR because_CS I_PPIS1 give_VV0 her_APPGE treat_NN1 food_NN1 ,_, but_CCB because_CS she_PPHS1 's_VBZ lonely._NNU 3_MC I_PPIS1 'm_VBM her_PPHO1 one_MC1 friend_NN1 now_RT ,_, the_AT only_JJ living_JJ being_NN1 she_PPHS1 gets_VVZ to_II see._NNU 4_MC By_II the_AT end_NN1 of_IO the_AT week_NNT1 ,_, I_PPIS1 rubberband_VV0 the_AT treat_NN1 food_NN1 dish_NN1 onto_II the_AT end_NN1 of_IO an_AT1 extra_JJ perch_NN and_CC put_VV0 it_PPH1 into_II the_AT cage_NN1 through_II the_AT door._NNU 5_MC I_PPIS1 lock_VV0 the_AT door_NN1 open_VV0 with_IW a_AT1 paper_NN1 clip._NNU 6_MC At_RR21 first_RR22 ,_, Birdie_NP1 's_GE shy_JJ ,_, but_CCB then_RT she_PPHS1 jumps_VVZ onto_II the_AT perch_NN I_PPIS1 'm_VBM holding_VVG and_CC sidehops_VVZ over_RP to_II the_AT treat_NN1 dish._NNU 7_MC It_PPH1 's_VBZ terrific_JJ to_TO see_VVI her_PPHO1 without_IW the_AT bars_NN2 between_II us._NNU 8_MC She_PPHS1 sits_VVZ eating_VVG the_AT treat_NN1 food_NN1 at_II the_AT opening_NN1 to_II the_AT door_NN1 and_CC looking_VVG at_II me._NNU 9_MC How_RRQ does_VDZ she_PPHS1 know_VVI to_TO look_VVI into_II my_APPGE eyes_NN2 and_CC not_XX at_II the_AT huge_JJ finger_NN1 next_II21 to_II22 her._NNU 10_MC After_CS she_PPHS1 's_VBZ finished_VVN eating_NN1 ,_, she_PPHS1 retreats_VVZ to_II the_AT middle_NN1 of_IO the_AT perch._NNU 11_MC I_PPIS1 lift_VV0 it_PPH1 gently_RR to_TO give_VVI her_PPHO1 a_AT1 ride_NN1 and_CC a_AT1 feeling_NN1 the_AT perch_NN is_VBZ part_NN1 of_IO me_PPIO1 and_CC not_XX the_AT cage._NNU 12_MC She_PPHS1 shifts_VVZ her_APPGE body_NN1 and_CC flips_VVZ her_APPGE wings_NN2 to_TO keep_VVI balance_NN1 ,_, then_RT looks_VVZ at_II me_PPIO1 and_CC makes_VVZ a_AT1 new_JJ sound_NN1 ,_, like_II peeEP_NN1 ;_; very_RG sharp._NNU 13_MC She_PPHS1 jumps_VVZ off_II the_AT perch_NN to_II the_AT bottom_NN1 of_IO the_AT cage._NNU 14_MC take_VV0 out_RP the_AT perch_NN and_CC try_VV0 to_TO talk_VVI to_II her_PPHO1 but_CCB she_PPHS1 ignores_VVZ me._NNU 15_MC She_PPHS1 drinks_VVZ water._NNU 16_MC She_PPHS1 does_VDZ n't_XX look_VVI at_II me_PPIO1 again_RT till_CS she_PPHS1 's_VBZ wiped_VVN off_II her_APPGE beak_NN1 and_CC stretched_VVN both_DB2 wings_NN2 ,_, one_MC1 at_II a_AT1 time._NNU 17_MC She_PPHS1 uses_VVZ her_APPGE feet_NN2 to_TO help_VVI stretch_VVI the_AT wings._NNU 18_MC Then_RT ,_, she_PPHS1 gives_VVZ a_AT1 small_JJ qUeeEEP_NN1 ?_? ._. 19_MC Generally_RR ,_, Birdie_NP1 looks_VVZ at_II me_PPIO1 more_RRR with_IW her_APPGE right_JJ eye_NN1 than_CSN her_APPGE left._NNU 20_MC It_PPH1 does_VDZ n't_XX matter_VVI which_DDQ side_NN1 of_IO the_AT cage_NN1 I_MC1 stand._NNU 21_MC She_PPHS1 turns_VVZ so_RR she_PPHS1 can_VM see_VVI me_PPIO1 with_IW her_PPHO1 right_RR eye._NNU 22_MC Also_RR ,_, when_CS she_PPHS1 reaches_VVZ with_IW her_APPGE foot_NN1 to_TO hold_VVI the_AT treat_NN1 dish_NN1 ,_, or_CC even_RR her_APPGE regular_JJ food_NN1 dish_NN1 ,_, she_PPHS1 does_VDZ it_PPH1 with_IW her_PPHO1 right_RR foot._NNU 23_MC She_PPHS1 'd_VM be_VBI right-handed_RR if_CS she_PPHS1 had_VHD hands_NN2 ;_; she_PPHS1 's_VBZ right-footed_JJ or_CC right-sided._JJ 24_MC She_PPHS1 approaches_VVZ and_CC does_VDZ most_DAT things_NN2 from_II the_AT right_JJ side_NN1 ._. 
(_( William_NP1 Wharton_NP1 ,_, Birdy_NP1 ,_, Jonathan_NP1 Cape_NP1 ,_, 1979_MC ,_, p._NN1 47_MC )_) If_CS there_EX are_VBR orthographic_JJ paragraph_NN1 divisions_NN2 in_II the_AT original_JJ version_NN1 of_IO this_DD1 text_NN1 which_DDQ were_VBDR made_VVN for_IF the_AT sake_NN1 of_IO appearance_NN1 on_II the_AT page_NN1 ,_, then_RT we_PPIS2 have_VH0 little_DA1 hope_NN1 of_IO identifying_VVG such_DA divisions_NN2 in_II any_DD formal_JJ way_NN1 ._. 
What_DDQ kind_NN1 of_IO formal_JJ marks_NN2 ,_, if_CS any_DD ,_, would_VM we_PPIS2 expect_VVI to_TO find_VVI at_II the_AT beginning_NN1 of_IO a_AT1 new_JJ paragraph_NN1 ?_? 
The_AT markers_NN2 Longacre_NP1 (_( 1979_MC )_) identifies_VVZ in_II narrative_JJ discourse_NN1 are_VBR inevitably_RR adverbial_JJ expressions_NN2 indicating_VVG temporal_JJ sequence_NN1 ._. 
It_PPH1 may_VM be_VBI that_CST the_AT general_JJ class_NN1 of_IO adverbials_NN2 which_DDQ can_VM appear_VVI initially_RR in_II a_AT1 sentence_NN1 could_VM be_VBI taken_VVN as_CSA possible_JJ markers_NN2 of_IO '_GE topic-shift_NN1 '_GE ._. 
Quirk_NN1 et_RA21 al._RA22 (_( 1972_MC :_: ch._NNU 8_MC )_) provide_VV0 lists_NN2 of_IO such_DA adverbials_NN2 in_II31 terms_II32 of_II33 adjuncts_NN2 ,_, conjuncts_NN2 and_CC disjuncts_NN2 ._. 
In_II fact_NN1 ,_, extract_NN1 (_( 14_MC )_) begins_VVZ with_IW an_AT1 adverbial_JJ clause_NN1 in_II initial_JJ position_NN1 ._. 
There_EX are_VBR two_MC other_JJ points_NN2 in_II this_DD1 extract_NN1 ,_, sentences_NN2 4_MC and_CC 10_MC ,_, where_CS adverbial_JJ clauses_NN2 occur_VV0 in_II sentence-initial_JJ position_NN1 ._. 
There_EX are_VBR four_MC other_JJ points_NN2 where_RRQ adverbial_JJ expressions_NN2 occur_VV0 sentence-initially_RR ,_, sentences_NN2 6_MC (_( At_RR21 first_RR22 )_) ,_, 18_MC (_( Then_RT )_) ,_, I9_FO (_( Generally_RR )_) ,_, and_CC 22_MC (_( Also_RR )_) ._. 
This_DD1 would_VM give_VVI us_PPIO2 six_MC possible_JJ breaks_NN2 ,_, formally_RR marked_VVN ,_, in_II the_AT structure_NN1 of_IO the_AT piece_NN1 of_IO text_NN1 ._. 
The_AT next_MD question_NN1 is_VBZ do_VD0 all_RR these_DD2 adverbial_JJ expressions_NN2 function_VV0 in_II the_AT same_DA way_NN1 ?_? 
After_II all_DB ,_, we_PPIS2 would_VM like_VVI to_TO distinguish_VVI between_II adverbials_NN2 which_DDQ indicate_VV0 a_AT1 connection_NN1 between_II one_MC1 sentence_NN1 and_CC the_AT next_MD and_CC those_DD2 adverbials_NN2 used_VMK to_TO link_VVI a_AT1 set_NN1 of_IO sentences_NN2 to_II another_DD1 set_NN1 ._. 
The_AT use_NN1 of_IO then_RT in_II 18_MC seems_VVZ to_TO introduce_VVI a_AT1 final_JJ action_NN1 in_II a_AT1 temporal_JJ sequence_NN1 of_IO actions_NN2 ._. 
We_PPIS2 can_VM conceive_VVI of_IO this_DD1 one_MC1 sentence_NN1 being_VBG separated_VVN from_II the_AT previous_JJ set_NN1 as_II a_AT1 form_NN1 of_IO distinct_JJ climax_NN1 ._. 
We_PPIS2 might_VM expect_VVI ,_, however_RR ,_, that_CST it_PPH1 would_VM more_RGR typically_RR occur_VVI as_II the_AT final_JJ sentence_NN1 of_IO a_AT1 paragraph_NN1 ,_, not_XX as_II a_AT1 climax_NN1 ,_, but_CCB as_CSA describing_VVG an_AT1 action_NN1 which_DDQ culminates_VVZ a_AT1 series_NN of_IO actions_NN2 ._. 
It_PPH1 is_VBZ followed_VVN by_II a_AT1 sentence_NN1 which_DDQ does_VDZ not_XX continue_VVI the_AT series_NN of_IO actions_NN2 and_CC which_DDQ begins_VVZ with_IW what_DDQ Quirk_NN1 et_RA21 al._RA22 (_( 1972_MC :_: 509_MC )_) would_VM characterise_VVI as_II a_AT1 '_GE style_NN1 disjunct_JJ '_GE ._. 
This_DD1 use_NN1 of_IO generally_RR ,_, in_II I9_FO ,_, effectively_RR separates_VVZ the_AT previous_JJ set_NN1 of_IO sentences_NN2 from_II the_AT next_MD set_NN1 describing_VVG a_AT1 particular_JJ habit_NN1 of_IO the_AT individual_NN1 involved_VVD ._. 
Within_II this_DD1 latter_DA set_NN1 ,_, one_MC1 sentence_NN1 begins_VVZ with_IW the_AT additive_JJ adjunct_NN1 ,_, also_RR ,_, in_II 22_MC ,_, which_DDQ could_VM be_VBI indicating_VVG that_CST there_EX are_VBR two_MC parts_NN2 to_II this_DD1 set_NN1 ._. 
It_PPH1 is_VBZ more_RGR likely_JJ that_CST the_AT sentence_NN1 beginning_VVG with_IW also_RR is_VBZ adding_VVG more_DAR detail_NN1 to_TO support_VVI the_AT general_JJ conclusion_NN1 that_CST the_AT individual_JJ concerned_JJ is_VBZ right-sided_JJ and_CC is_VBZ part_NN1 of_IO the_AT internal_JJ structure_NN1 of_IO a_AT1 paragraph_NN1 beginning_VVG with_IW Generally_RR ._. 
The_AT other_JJ adverbial_JJ ,_, at_RR21 first_RR22 ,_, in_II 6_MC ,_, seems_VVZ to_TO be_VBI part_NN1 of_IO a_AT1 sentence-internal_JJ construction_NN1 ,_, especially_RR when_CS we_PPIS2 see_VV0 the_AT then_RT which_DDQ follows_VVZ ._. 
The_AT events_NN2 described_VVN in_II this_DD1 sentence_NN1 fall_NN1 within_II the_AT set_NN1 of_IO events_NN2 described_VVN as_II happening_VVG by_II the_AT end_NN1 of_IO the_AT week_NNT1 (_( in_II 4_MC )_) ._. 
Thus_RR ,_, we_PPIS2 have_VH0 reduced_VVN the_AT number_NN1 of_IO possible_JJ breaks_NN2 in_II this_DD1 text_NN1 to_II three_MC ,_, so_CS21 that_CS22 we_PPIS2 can_VM suggest_VVI that_CST there_EX are_VBR four_MC paragraphs_NN2 ,_, beginning_VVG at_II sentences_NN2 1_MC1 ,_, 4_MC ,_, 10_MC and_CC 19_MC ._. 
The_AT reader_NN1 may_VM suggest_VVI other_JJ possible_JJ breaks_NN2 ,_, as_CSA ,_, for_REX21 example_REX22 ,_, in_II 9_MC ,_, where_RRQ there_EX is_VBZ a_AT1 sentence_NN1 structure_NN1 (_( an_AT1 interrogative_NN1 )_) quite_RG different_JJ from_II the_AT structure_NN1 of_IO the_AT rest_NN1 of_IO the_AT text_NN1 sentences_NN2 ._. 
An_AT1 argument_NN1 for_IF a_AT1 break_NN1 here_RL would_VM seem_VVI quite_RG reasonable_JJ since_CS this_DD1 sentence_NN1 is_VBZ structurally_RR marked_VVN as_CSA separate_JJ ._. 
No_RR21 doubt_RR22 the_AT reader_NN1 could_VM also_RR think_VVI of_IO an_AT1 argument_NN1 ,_, mainly_RR in_II stylistic_JJ terms_NN2 ,_, for_IF treating_VVG this_DD1 sentence_NN1 as_II part_NN1 of_IO the_AT preceding_JJ set_NN1 ._. 
It_PPH1 may_VM be_VBI the_AT case_NN1 that_CST ,_, taking_VVG stylistic_JJ considerations_NN2 more_RGR generally_RR ,_, the_AT reader_NN1 would_VM wish_VVI to_TO divide_VVI this_DD1 text_NN1 into_II separate_JJ paragraphs_NN2 at_II points_NN2 where_RRQ there_EX are_VBR no_AT formal_JJ markers_NN2 at_RR21 all_RR22 ._. 
We_PPIS2 would_VM assume_VVI that_CST the_AT discussion_NN1 ,_, in_II such_DA a_AT1 case_NN1 ,_, would_VM cease_VVI to_TO be_VBI a_AT1 discussion_NN1 which_DDQ appealed_VVD to_II primarily_RR linguistic_JJ evidence_NN1 in_II this_DD1 piece_NN1 of_IO discourse_NN1 ._. 
On_II the_AT basis_NN1 of_IO some_DD formal_JJ linguistic_JJ markers_NN2 ,_, we_PPIS2 have_VH0 suggested_VVN that_CST there_EX are_VBR four_MC paragraphs_NN2 in_II extract_NN1 (_( 14_MC )_) ._. 
We_PPIS2 may_VM have_VHI been_VBN led_VVN to_II finding_VVG those_DD2 four_MC paragraphs_NN2 because_CS they_PPHS2 are_VBR ,_, in_II fact_NN1 ,_, the_AT divisions_NN2 which_DDQ actually_RR appear_VV0 in_II the_AT original_JJ and_CC we_PPIS2 merely_RR sought_VVN additional_JJ evidence_NN1 to_TO support_VVI the_AT way_NN1 the_AT author_NN1 had_VHD divided_VVN up_RP his_APPGE discourse_NN1 ._. 
Yet_RR this_DD1 point_NN1 highlights_VVZ the_AT fact_NN1 that_CST the_AT exercise_NN1 we_PPIS2 have_VH0 performed_VVN on_II extract_NN1 (_( 14_MC )_) was_VBDZ an_AT1 extremely_RR artificial_JJ treatment_NN1 of_IO written_JJ discourse_NN1 ._. 
We_PPIS2 began_VVD by_II removing_VVG one_MC1 of_IO the_AT primary_JJ indicators_NN2 of_IO '_GE topic-shift_NN1 '_GE available_JJ to_II a_AT1 writer_NN1 ,_, that_DD1 of_IO indenting_VVG a_AT1 line_NN1 in_II his_APPGE text_NN1 ._. 
Rather_CS21 than_CS22 treat_VVI the_AT indenting_NN1 of_IO the_AT first_MD line_NN1 of_IO a_AT1 paragraph_NN1 as_CSA simply_RR some_DD cosmetic_JJ device_NN1 ,_, as_CSA Longacre_NP1 (_( 1979_MC )_) does_VDZ ,_, we_PPIS2 might_VM look_VVI upon_II it_PPH1 as_II an_AT1 indication_NN1 by_II a_AT1 writer_NN1 of_IO what_DDQ he_PPHS1 intends_VVZ us_PPIO2 to_TO treat_VVI as_II the_AT beginning_NN1 of_IO a_AT1 new_JJ part_NN1 of_IO his_APPGE text_NN1 ._. 
If_CS the_AT writer_NN1 also_RR uses_VVZ adverbial_JJ expressions_NN2 initially_RR in_II the_AT first_MD sentence_NN1 of_IO this_DD1 new_JJ part_NN1 of_IO his_APPGE text_NN1 ,_, then_RT we_PPIS2 might_VM say_VVI we_PPIS2 have_VH0 overwhelming_JJ evidence_NN1 that_CST the_AT writer_NN1 is_VBZ marking_VVG a_AT1 '_GE topic-shift_NN1 '_GE in_II his_APPGE discourse_NN1 ._. 
We_PPIS2 are_VBR ,_, after_II all_DB ,_, performing_VVG a_AT1 descriptive_JJ and_CC not_XX a_AT1 prescriptive_JJ exercise_NN1 when_CS we_PPIS2 undertake_VV0 discourse_NN1 analysis_NN1 ._. 
We_PPIS2 do_VD0 not_XX wish_VVI to_TO say_VVI how_RRQ a_AT1 writer_NN1 should_VM organise_VVI his_APPGE written_JJ discourse_NN1 into_II paragraphs_NN2 before_CS we_PPIS2 have_VH0 managed_VVN to_TO characterise_VVI ,_, in_II any_DD comprehensive_JJ way_NN1 ,_, how_RRQ writers_NN2 typically_RR do_VD0 so_RR ._. 
The_AT investigation_NN1 of_IO what_DDQ writers_NN2 typically_RR do_VD0 when_CS marking_VVG the_AT structure_NN1 of_IO their_APPGE texts_NN2 would_VM seem_VVI to_TO be_VBI a_AT1 more_RGR appropriate_JJ goal_NN1 of_IO discourse_NN1 analysis_NN1 ._. 
For_REX21 example_REX22 ,_, rather_CS21 than_CS22 dismiss_VVI the_AT orthographic_JJ paragraph_NN1 format_NN1 to_TO be_VBI found_VVN in_II newspaper_NN1 articles_NN2 as_CSA ,_, in_II some_DD way_NN1 ,_, a_AT1 deviation_NN1 from_II the_AT '_GE true_JJ '_GE paragraph_NN1 structure_NN1 of_IO what_DDQ is_VBZ being_VBG written_VVN ,_, it_PPH1 would_VM be_VBI more_RGR appropriate_JJ for_IF discourse_NN1 analysts_NN2 to_TO describe_VVI the_AT journalistic_JJ format_NN1 as_II one_MC1 form_NN1 of_IO written_JJ discourse_NN1 organisation_NN1 ._. 
The_AT paragraph_NN1 structure_NN1 of_IO different_JJ genres_NN2 ,_, such_II21 as_II22 scientific_JJ textbook_NN1 writing_NN1 ,_, repair_VV0 manuals_NN2 ,_, nineteenth-century_JJ novels_NN2 ,_, etc._RA could_VM then_RT be_VBI characterised_VVN ,_, and_CC statements_NN2 could_VM be_VBI made_VVN about_RP ,_, for_REX21 example_REX22 ,_, the_AT '_GE norms_NN2 '_GE or_CC regular_JJ features_NN2 of_IO topic-shift_NN1 in_II such_DA genres_NN2 ._. 
On_II the_AT basis_NN1 of_IO such_DA genre-specific_JJ descriptions_NN2 of_IO '_GE topic-shift_NN1 '_GE markers_NN2 ,_, it_PPH1 should_VM be_VBI possible_JJ to_TO make_VVI linguistic_JJ ,_, as_II31 opposed_II32 to_II33 literary_JJ ,_, statements_NN2 about_II the_AT structure_NN1 of_IO English_JJ written_JJ discourse_NN1 which_DDQ reflect_VV0 the_AT writer_NN1 's_GE purpose_NN1 ._. 
Thus_RR ,_, in_II producing_VVG a_AT1 narrative_NN1 ,_, the_AT writer_NN1 must_VM provide_VVI some_DD indications_NN2 of_IO change_NN1 of_IO time_NNT1 and_CC place_NN1 ,_, as_CSA Grimes_NP1 (_( 1975_MC :_: 102_MC )_) has_VHZ pointed_VVN out_RP ._. 
In_II presenting_VVG a_AT1 philosophical_JJ argument_NN1 ,_, however_RR ,_, the_AT writer_NN1 can_VM range_VVI over_RP different_JJ times_NNT2 and_CC places_NN2 within_II a_AT1 single_JJ paragraph_NN1 ,_, but_CCB must_VM mark_VVI out_RP changes_NN2 in_II the_AT direction_NN1 of_IO his_APPGE argument_NN1 ._. 
Taking_VVG a_AT1 random_JJ page_NN1 from_II the_AT writings_NN2 of_IO Karl_NP1 Popper_NN1 ,_, one_PN1 can_VM see_VVI the_AT structure_NN1 of_IO the_AT discourse_NN1 in_II skeleton_NN1 form_NN1 by_II taking_VVG the_AT first_MD phrase_NN1 or_CC sentence_NN1 of_IO each_DD1 paragraph._NNU (_( 15_MC )_) para_NN1 1_MC1 :_: Other_JJ questions_NN2 have_VH0 sometimes_RT been_VBN asked_VVN ..._... para_NN1 2_MC :_: Another_DD1 question_NN1 sometimes_RT asked_VVN is_VBZ this_DD1 ..._... para_NN1 3_MC :_: The_AT only_JJ correct_JJ answer_NN1 is_VBZ the_AT straightforward_JJ one_PN1 ..._... para_NN1 4_MC :_: It_PPH1 has_VHZ also_RR been_VBN said_VVN that_CST the_AT problem_NN1 of_IO induction_NN1 is_VBZ ..._... 
(_( Popper_NN1 ,_, 1963_MC :_: 56_MC )_) Eventually_RR ,_, it_PPH1 should_VM also_RR be_VBI possible_JJ to_TO specify_VVI those_DD2 markers_NN2 of_IO '_GE topic-shift_NN1 '_GE which_DDQ occur_VV0 in_II all_DB forms_NN2 of_IO written_JJ discourse_NN1 ._. 
We_PPIS2 might_VM find_VVI that_CST it_PPH1 is_VBZ indeed_RR the_AT case_NN1 that_CST the_AT use_NN1 of_IO '_GE But_CCB '_GE at_II the_AT beginning_NN1 of_IO a_AT1 paragraph_NN1 as_CSA described_VVN by_II van_NP1 Dijk_NP1 (_( 1977_MC :_: 139_MC )_) ,_, is_VBZ a_AT1 very_RG general_JJ marker_NN1 of_IO topic_NN1 change_NN1 ._. 
Other_JJ examples_NN2 of_IO what_DDQ van_NP1 Dijk_NP1 (_( 1977_MC :_: 150_MC )_) terms_NN2 macro-structure_NN1 connectives_NN2 are_VBR '_GE furthermore_RR '_GE ,_, '_GE however_RR '_GE ,_, and_CC '_GE so_RR '_GE ._. 
We_PPIS2 shall_VM discuss_VVI the_AT concept_NN1 of_IO macro-structures_NN2 in_II discourse_NN1 in_II section_NN1 3._MC 7_MC on_II the_AT proposition-based_JJ analysis_NN1 of_IO discourse_NN1 ._. 
Paratones_VVZ So_RG far_RR we_PPIS2 have_VH0 concentrated_VVN on_II structural_JJ markers_NN2 in_II written_JJ discourse_NN1 ._. 
In_II spoken_JJ discourse_NN1 ,_, there_EX is_VBZ not_XX the_AT visual_JJ prompt_NN1 of_IO paragraph-initial_JJ line_NN1 indentation_NN1 to_TO indicate_VVI a_AT1 division_NN1 in_II the_AT discourse_NN1 structure_NN1 ._. 
How_RRQ do_VD0 speakers_NN2 mark_VVI '_GE topic-shifts_NN2 '_GE ?_? 
One_MC1 suggestion_NN1 is_VBZ that_CST there_EX are_VBR ,_, in_II fact_NN1 ,_, structural_JJ units_NN2 of_IO spoken_JJ discourse_NN1 which_DDQ take_VV0 the_AT form_NN1 of_IO '_GE speech_NN1 paragraphs_NN2 '_GE and_CC have_VH0 been_VBN called_VVN paratones_NN2 (_( see_VV0 Brown_NP1 ,_, 1977_MC :_: 86_MC )_) ._. 
Some_DD support_NN1 for_IF the_AT notion_NN1 that_CST there_EX are_VBR ways_NN2 of_IO marking_VVG the_AT boundaries_NN2 of_IO '_GE speech_NN1 paragraphs_NN2 '_GE can_VM be_VBI found_VVN in_II a_AT1 common_JJ practice_NN1 of_IO people_NN who_PNQS are_VBR asked_VVN to_TO read_VVI pieces_NN2 of_IO written_JJ text_NN1 aloud_RR ._. 
They_PPHS2 use_VV0 intonational_JJ cues_NN2 to_TO signal_VVI the_AT start_NN1 of_IO a_AT1 new_JJ paragraph_NN1 ._. 
The_AT '_GE speech_NN1 paragraph_NN1 '_GE ,_, or_CC paratone_NN1 ,_, like_II the_AT orthographic_JJ paragraph_NN1 ,_, is_VBZ identified_VVN by_II its_APPGE boundary_NN1 markers_NN2 ._. 
The_AT marking_NN1 of_IO the_AT start_NN1 of_IO a_AT1 paratone_NN1 ,_, then_RT ,_, is_VBZ clearly_RR one_MC1 device_NN1 which_DDQ speakers_NN2 can_VM use_VVI to_TO indicate_VVI a_AT1 topic-shift_NN1 ._. 
Since_CS the_AT paratone_NN1 is_VBZ a_AT1 much_RR less_RGR familiar_JJ concept_NN1 than_CSN the_AT orthographic_JJ paragraph_NN1 ,_, it_PPH1 may_VM be_VBI useful_JJ to_TO have_VHI its_APPGE identifying_JJ features_NN2 described_VVN ._. 
At_II the_AT beginning_NN1 of_IO a_AT1 paratone_NN1 ,_, the_AT speaker_NN1 typically_RR uses_VVZ an_AT1 introductory_JJ expression_NN1 to_TO announce_VVI what_DDQ he_PPHS1 specifically_RR intends_VVZ to_TO talk_VVI about_II ._. 
This_DD1 introductory_JJ expression_NN1 is_VBZ made_VVN phonologically_RR prominent_JJ and_CC the_AT whole_NN1 of_IO the_AT first_MD clause_NN1 or_CC sentence_NN1 in_II a_AT1 paratone_NN1 may_VM be_VBI uttered_VVN with_IW raised_JJ pitch_NN1 ._. 
The_AT end_NN1 of_IO a_AT1 paratone_NN1 is_VBZ marked_VVN in_II a_AT1 way_NN1 similar_JJ to_II the_AT '_GE turn_NN1 signal_NN1 '_GE discussed_VVN by_II those_DD2 who_PNQS investigate_VV0 conversational_JJ discourse_NN1 as_II a_AT1 process_NN1 of_IO social_JJ interaction_NN1 (_( cf._VV0 Duncan_NP1 ,_, 1974_MC ;_; Sacks_NP1 et_RA21 al._RA22 ,_, 1974_MC )_) ._. 
It_PPH1 can_VM be_VBI marked_VVN by_II very_RG low_JJ pitch_NN1 ,_, even_RR on_II lexical_JJ items_NN2 ,_, loss_NN1 of_IO amplitude_NN1 and_CC a_AT1 lengthy_JJ pause_NN1 ._. 
Alternatively_RR ,_, the_AT speaker_NN1 can_VM use_VVI a_AT1 summarising_JJ phrase_NN1 ,_, often_RR repeating_VVG the_AT introductory_JJ expression_NN1 ,_, not_XX necessarily_RR low_JJ in_II pitch_NN1 ,_, but_CCB also_RR followed_VVN by_II a_AT1 lengthy_JJ pause_NN1 ._. 
The_AT most_RGT consistent_JJ paratone-final_JJ marker_NN1 is_VBZ the_AT long_JJ pause_NN1 ,_, normally_RR exceeding_VVG one_MC1 second_NNT1 ._. 
We_PPIS2 shall_VM examine_VVI an_AT1 extract_NN1 from_II conversational_JJ discourse_NN1 containing_VVG a_AT1 longish_JJ paratone_NN1 which_DDQ illustrates_VVZ the_AT features_NN2 just_RR described_VVN ._. 
It_PPH1 is_VBZ relevant_JJ that_CST the_AT topic_NN1 framework_NN1 for_IF this_DD1 extract_NN1 (_( 16_MC )_) should_VM contain_VVI information_NN1 about_II the_AT speaker_NN1 (_( female_NN1 ,_, 20+_FO ,_, Edinburgh-resident_NN1 )_) and_CC the_AT preceding_JJ discourse_NN1 (_( the_AT types_NN2 of_IO drinks_NN2 the_AT participants_NN2 had_VHD encountered_VVN in_II different_JJ types_NN2 of_IO bars_NN2 during_II their_APPGE respective_JJ recent_JJ holidays_NN2 in_II the_AT United_NP1 States_NP1 )_) ._. 
It_PPH1 is_VBZ also_RR worth_II noting_VVG that_CST in_II Edinburgh_NP1 Scottish_JJ English_NN1 ,_, phonologically_RR prominent_JJ syllables_NN2 are_VBR typically_RR uttered_VVN with_IW raised_JJ or_CC high_JJ pitch_NN1 and_CC need_VV0 not_XX have_VHI the_AT type_NN1 of_IO pitch_NN1 movement_NN1 associated_VVN with_IW phonological_JJ prominence_NN1 in_II descriptions_NN2 of_IO standard_JJ southern_JJ English_NN1 (_( cf._VV0 Brown_NP1 et_RA21 al._RA22 ,_, 1980_MC )_) ._. 
(_( For_IF an_AT1 explanation_NN1 of_IO the_AT stave_NN1 representation_NN1 of_IO intonation_NN1 used_VVD ,_, see_VV0 the_AT '_GE transcription_NN1 conventions_NN2 '_GE on_II p._NNU xii._MC )_) (_( 16_MC )_) I_PPIS1 found_VVD my_APPGE drink_NN1 was_VBDZ a_AT1 great_JJ problem_NN1 with_IW them_PPHO2 because_CS (_( )_) at_II that_DD1 time_NNT1 I_PPIS1 drank_VVD whisky_NN1 and_CC lemonade_NN1 +_FO and_CC I_PPIS1 would_VM (_( )_) go_VV0 and_CC ask_VV0 for_IF whisky_NN1 and_CC lemonade_NN1 and_CC I_PPIS1 would_VM get_VVI (_( )_) whisky_NN1 and_CC lemon_NN1 +_FO because_CS you_PPY have_VH0 to_TO ask_VVI for_IF whisky_NN1 (_( )_) or_CC scotch_NN1 and_CC seven_MC up_II +_FO you_PPY know_VV0 +_FO I_ZZ1 eventually_RR (_( )_) cottoned_VVD on_RP to_II it_PPH1 +_FO but_CCB +_FO and_CC they_PPHS2 could_VM n't_XX get_VVI over_RP (_( )_) the_AT fact_NN1 that_CST I_PPIS1 did_VDD n't_XX like_VVI ice_NN1 in_II whisky_NN1 and_CC of_RR21 course_RR22 (_( )_) they_PPHS2 either_RR gave_VVD me_PPIO1 ice_NN1 whether_CSW I_PPIS1 wanted_VVD it_PPH1 or_CC not_XX or_CC (_( )_) they_PPHS2 stacked_VVD the_AT glass_NN1 up_II +_FO right_RR up_II21 to_II22 the_AT level_NN1 that_CST (_( )_) you_PPY would_VM normally_RR have_VHI if_CSW you_PPY had_VHD ice_NN1 in_II your_APPGE drink_NN1 (_( )_) anyway_RR +_FO and_CC consequently_RR I_PPIS1 got_VVD ploughed_JJ +_FO frequently_RR +_FO (_( )_) and_CC that_DD1 's_VBZ that_DD1 I_ZZ1 +_FO I_ZZ1 tended_VVD to_TO stick_VVI to_II my_APPGE drink_NN1 ++_FO (_( )_) This_DD1 paratone_NN1 begins_VVZ with_IW an_AT1 introductory_JJ expression_NN1 my_APPGE drink_NN1 ,_, uttered_VVD very_RG high_JJ in_II the_AT pitch_NN1 range_NN1 ,_, and_CC closes_VVZ with_IW the_AT same_DA expression_NN1 ,_, low_JJ in_II the_AT pitch_NN1 range_NN1 ,_, as_II part_NN1 of_IO the_AT speaker_NN1 's_GE summing-up_NN1 ._. 
The_AT internal_JJ pauses_NN2 are_VBR brief_JJ ,_, none_PN exceeding_VVG 0.5_MC seconds_NNT2 ,_, but_CCB the_AT final_JJ pause_NN1 marking_VVG the_AT end_NN1 of_IO the_AT paratone_NN1 is_VBZ long_RR (_( 1.6_MC seconds_NNT2 )_) ._. 
Those_DD2 are_VBR the_AT formal_JJ markers_NN2 of_IO the_AT boundaries_NN2 of_IO this_DD1 paratone_NN1 ._. 
Of_RR21 course_RR22 ,_, there_EX are_VBR internal_JJ aspects_NN2 ,_, such_II21 as_II22 the_AT semantic_JJ cohesion_NN1 within_II the_AT lexical_JJ field_NN1 established_VVN by_II my_APPGE drink_NN1 ,_, which_DDQ could_VM also_RR be_VBI appealed_VVN to_II in_II claiming_VVG that_CST this_DD1 chunk_NN1 of_IO discourse_NN1 is_VBZ a_AT1 unit_NN1 of_IO some_DD kind_NN1 ._. 
However_RR ,_, this_DD1 type_NN1 of_IO internal_JJ cohesion_NN1 is_VBZ not_XX a_AT1 necessary_JJ feature_NN1 of_IO the_AT structural_JJ unit_NN1 we_PPIS2 have_VH0 described_VVN as_II the_AT paratone_NN1 ._. 
It_PPH1 might_VM be_VBI argued_VVN that_CST there_EX are_VBR two_MC paratones_NN2 ,_, and_CC not_XX one_PN1 ,_, in_II this_DD1 extract_NN1 ._. 
There_EX appears_VVZ to_TO be_VBI a_AT1 break_NN1 where+_FO but_CCB +_FO is_VBZ used_VVN ._. 
Indeed_RR ,_, just_RR prior_II21 to_II22 but_CCB ,_, there_EX is_VBZ what_DDQ has_VHZ been_VBN described_VVN as_II a_AT1 '_GE possible_JJ completion_NN1 point_NN1 '_GE ._. 
The_AT speaker_NN1 has_VHZ come_VVN to_II the_AT end_NN1 of_IO a_AT1 sentence_NN1 and_CC pauses_NN2 ._. 
It_PPH1 is_VBZ a_AT1 point_NN1 at_II which_DDQ those_DD2 who_PNQS analyse_VV0 conversation_NN1 in_II31 terms_II32 of_II33 '_GE turn-taking_NN1 '_GE (_( Sacks_NP1 et_RA21 al._RA22 ,_, 1974_MC )_) would_VM suggest_VVI that_CST another_DD1 speaker_NN1 could_VM take_VVI over_II the_AT turn_NN1 ._. 
However_RR ,_, the_AT speaker_NN1 in_II this_DD1 extract_NN1 immediately_RR produces_VVZ an_AT1 '_GE utterance_NN1 incompletor_NN1 '_GE in_II this_DD1 case_NN1 but_CCB ,_, though_CS any_DD clause_NN1 connector_NN1 would_VM do_VDI making_NN1 ,_, as_CSA Coulthard_NP1 (_( 1977_MC :_: 56_MC )_) points_VVZ out_RP ,_, a_AT1 potentially_RR complete_JJ utterance_NN1 into_II an_AT1 incomplete_JJ one_PN1 ._. 
After_CS another_DD1 brief_JJ pause_NN1 ,_, the_AT speaker_NN1 continues_VVZ ,_, using_VVG and_CC to_TO indicate_VVI that_DD1 what_DDQ she_PPHS1 is_VBZ going_VVGK to_TO say_VVI is_VBZ connected_VVN to_II what_DDQ she_PPHS1 has_VHZ just_RR said_VVN ._. 
We_PPIS2 would_VM not_XX want_VVI to_TO describe_VVI this_DD1 possible_JJ completion_NN1 point_NN1 (_( or_CC any_DD other_JJ which_DDQ occurs_VVZ in_II this_DD1 extract_NN1 )_) as_II a_AT1 paratone-boundary_NN1 ._. 
The_AT formal_JJ markers_NN2 ,_, low_JJ pitch_NN1 close_VV0 plus_II lengthy_JJ pause_NN1 plus_II raised_JJ pitch_NN1 introductory_JJ expression_NN1 ,_, are_VBR not_XX present_JJ ._. 
In_II intuitive_JJ terms_NN2 ,_, we_PPIS2 might_VM also_RR say_VVI that_DD1 what_DDQ follows_VVZ +_FO but_CCB +_FO is_VBZ not_XX on_II a_AT1 separate_JJ speaker_NN1 's_GE topic_NN1 ,_, but_CCB continues_VVZ the_AT talk_NN1 '_GE about_II '_GE 'my_JJ drink'_NN1 ._. 
At_II the_AT end_NN1 of_IO this_DD1 extract_NN1 ,_, there_EX is_VBZ an_AT1 obvious_58 '_GE completion_NN1 point_NN1 '_GE ._. 
In_II '_GE turn-taking_NN1 '_GE terms_NN2 ,_, it_PPH1 is_VBZ a_AT1 point_NN1 at_II which_DDQ another_DD1 speaker_NN1 is_VBZ free_JJ to_TO take_VVI over_RP ._. 
However_RR ,_, in_II this_DD1 part_NN1 of_IO the_AT conversation_NN1 ,_, one_MC1 speaker_NN1 clearly_RR '_GE has_VHZ the_AT floor_NN1 '_GE and_CC she_PPHS1 is_VBZ allowed_VVN to_TO continue_VVI ,_, as_CSA shown_VVN in_II extract_NN1 (_( I_ZZ1 7_MC )_) ._. 
(_( 17_MC )_) oh_UH apart_II21 from_II22 once_RR when_CS we_PPIS2 went_VVD we_PPIS2 found_VVD em_FU +_FO an_AT1 Irish_JJ bar_NN1 (_( )_) in_II San_NP1 Francisco_NP1 that_DD1 was_VBDZ famous_JJ for_IF its_APPGE Irish_JJ coffees_NN2 +_FO (_( )_) In_II beginning_VVG a_AT1 new_JJ paratone_NN1 ,_, the_AT speaker_NN1 marks_NN2 as_CSA intonationally_RR prominent_JJ two_MC expressions_NN2 an_AT1 Irish_JJ bar_NN1 and_CC Irish_JJ coffees_NN2 ._. 
In_II the_AT course_NN1 of_IO the_AT paratone_NN1 ,_, she_PPHS1 talks_VVZ about_II both_DB2 the_AT bar_NN1 and_CC the_AT Irish_JJ coffee_NN1 made_VVD there_RL ._. 
It_PPH1 seems_VVZ quite_RG reasonable_JJ to_TO assume_VVI that_DD1 ,_, when_CS a_AT1 speaker_NN1 is_VBZ organising_VVG a_AT1 '_GE speech_NN1 paragraph_NN1 '_GE which_DDQ has_VHZ two_MC connected_JJ elements_NN2 as_CSA its_APPGE foci_NN2 ,_, both_DB2 elements_NN2 can_VM be_VBI made_VVN phonologically_RR prominent_JJ in_II the_AT introduction_NN1 ._. 
When_CS the_AT speaker_NN1 closes_VVZ this_DD1 paratone_NN1 ,_, she_PPHS1 repeats_VVZ one_MC1 of_IO her_APPGE introductory_JJ expressions_NN2 it_PPH1 was_VBDZ very_RG good_JJ Irish_JJ coffee_NN1 too_RR not_XX particularly_RR low_JJ in_II the_AT pitch_NN1 range_NN1 ,_, but_CCB followed_VVD by_II a_AT1 lengthy_JJ pause_NN1 ._. 
Some_DD of_IO the_AT features_NN2 we_PPIS2 have_VH0 described_VVN as_II marking_VVG paratone_NN1 boundaries_NN2 in_II spoken_JJ discourse_NN1 can_VM ,_, of_RR21 course_RR22 ,_, have_VH0 other_JJ functions_NN2 ._. 
Although_CS the_AT lengthy_JJ pause_NN1 is_VBZ also_RR identified_VVN by_II Chafe_NP1 (_( 1979_MC :_: 176_MC )_) as_II an_AT1 indication_NN1 of_IO segmentation_NN1 in_II his_APPGE spoken_JJ discourse_NN1 data_NN comparable_JJ to_II paragraphing_VVG in_II written_JJ discourse_NN1 ,_, the_AT intonational_JJ features_NN2 we_PPIS2 appealed_VVD to_II can_VM have_VHI other_JJ ,_, quite_RG different_JJ ,_, functions_NN2 ._. 
Some_DD of_IO these_DD2 we_PPIS2 will_VM discuss_VVI in_II detail_NN1 in_II Chapter_NN1 5_MC ._. 
What_DDQ we_PPIS2 have_VH0 described_VVN is_VBZ the_AT use_NN1 of_IO the_AT combination_NN1 of_IO these_DD2 formal_JJ markers_NN2 by_II speakers_NN2 to_TO indicate_VVI a_AT1 shift_NN1 in_II what_DDQ they_PPHS2 're_VBR talking_VVG about_II ._. 
There_EX may_VM be_VBI other_JJ ,_, more_RGR subtle_JJ ,_, indicators_NN2 of_IO topic-shift_NN1 used_VVN by_II conversationalists_NN2 which_DDQ we_PPIS2 have_VH0 ignored_VVN ._. 
The_AT significance_NN1 of_IO '_GE speaker_NN1 gaze_NN1 '_GE ,_, as_CSA described_VVN by_II Kendon_NP1 (_( 1967_MC )_) and_CC specific_JJ '_GE body_NN1 movements_NN2 '_GE (_( de_FW Long_RR ,_, 1974_MC )_) in_II signalling_VVG speaker_NN1 change_NN1 in_II conversation_NN1 may_VM also_RR be_VBI relevant_JJ in_II topic_NN1 change_NN1 ._. 
The_AT occurrence_NN1 of_IO different_JJ types_NN2 of_IO '_GE fillers_NN2 '_GE such_II21 as_II22 '_GE well_NN1 '_GE ,_, '_" mmm_UH '_GE ,_, '_" you_PPY know_VV0 '_GE ,_, '_" er_FU '_GE ,_, and_CC others_NN2 may_VM also_RR regularly_RR coincide_VVI with_IW topic-shifts_NN2 ._. 
We_PPIS2 have_VH0 concentrated_VVN ,_, however_RR ,_, on_II some_DD of_IO the_AT primary_NN1 ,_, easily_RR identifiable_JJ formal_JJ markers_NN2 used_VVN by_II writers_NN2 and_CC speakers_NN2 to_TO indicate_VVI structural_JJ divisions_NN2 in_II the_AT discourse_NN1 they_PPHS2 produce_VV0 ._. 
We_PPIS2 emphasise_VV0 once_RR21 again_RR22 that_DD1 ,_, although_CS we_PPIS2 can_VM regularly_RR identify_VVI such_DA structural_JJ markers_NN2 ,_, their_APPGE appearance_NN1 in_II discourse_NN1 should_VM not_XX be_VBI treated_VVN in_II any_DD way_NN1 as_58 '_GE rule-governed_NN1 '_GE ._. 
They_PPHS2 represent_VV0 optional_JJ cues_NN2 which_DDQ writers_NN2 and_CC speakers_NN2 may_VM use_VVI in_II organising_VVG what_DDQ they_PPHS2 want_VV0 to_TO communicate_VVI ._. 
Failure_NN1 to_TO mark_VVI out_RP explicitly_RR the_AT structural_JJ organisation_NN1 of_IO what_DDQ a_AT1 speaker_NN1 wishes_VVZ to_TO communicate_VVI may_VM make_VVI the_AT addressee_NN1 's_GE task_NN1 of_IO interpretation_NN1 more_RGR difficult_JJ ,_, perhaps_RR ,_, but_CCB ,_, by_II itself_PPX1 ,_, would_VM not_XX necessarily_RR constitute_VVI a_AT1 failure_NN1 to_TO communicate_VVI ._. 
