{
    "paper_id": "J78-3017",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T03:04:12.890136Z"
    },
    "title": "BOUND VARIABLES AND OTHER ANAPHORS",
    "authors": [
        {
            "first": "Barbara",
            "middle": [
                "H"
            ],
            "last": "Partee",
            "suffix": "",
            "affiliation": {},
            "email": ""
        },
        {
            "first": "Ilniv",
            "middle": [],
            "last": "O F Mass",
            "suffix": "",
            "affiliation": {},
            "email": ""
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "",
    "pdf_parse": {
        "paper_id": "J78-3017",
        "_pdf_hash": "",
        "abstract": [],
        "body_text": [
            {
                "text": "When a noun phrase o r a pronoun occurs i n a sentence, i t i s frequently appropriate t o ask what e n t f t y i t refers to, but i t i s well known t h a t not a l l uses o f noun phrases and pronouns are referential i n t h i s simple sense. I n computatlon-a1 approaches t o language processing, I believe the main t h r u s t Jn t h i s area has been toward understanding those referential uses o f NP1s and pronouns which require the use o f both linguist i c and nbn-1 i n g u i s t i c inferences t o determine the most plausible referent f o r the expression. M y emphasis i n t h i s paper w i 11 be somewhat d i fferent. I be1 ieve that recent work by linguisds, logictans, and philosophers i s leading t o convergence on the view t h a t there are two fundamen: t a l l y df s t i n c t uses o f pronouns which have t o be treated quite separately: ( i ) a use t h a t corresponds t o the 1 ogician's use o f bwnd variables, and ( i t ) a use which I w i 11 c a l l , f o r want o f a better name, a pragmatic use. I t can be argued that bound variable pronouns are r e s t r i c t e d t o occurrences i n syntactic construction wS t h the1 r antecedents, and are fu1 l y interpreted a t the level lef semantics , whi le pragmatic pronouns need n o t have 1 ingui s t i c antecedents a t \"a1 1, and require pragmatics as well as semantics f o r t h e i r interpretation.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "The clearest cases o f bound variable anaphora invol ve antecedents 1i ke ever man and no man which are singular i n form -%but o not r e f e r t o individuals, as i n (1) and (2).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "( I ) Ever man put a screen i n f r o n t o f him.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "(2) & w i l l admit t h a the i s s l e e z When the he o f (2) i s understood as anaphorically re1 ated t a h e noun phrase no child, the he clearly does not r e f e r t o a particular inafvidual Rather, the sentence can be understood as the r e s u l t of binding an upen sentence, ( 3 ) , with a q u a n t i f i e r phrase, no child.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "(3) Heo w i l l admit t h a t heo i s sleepy.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "(It i s imnaterial f o r the purposes o f t h i s paper whether we view the process i n question as a enerative one, as i n Montague (1973) o r Lakoff 9 1971) or as an interpretive one, as I n Jackendoff (1972) o r the I-gramnar Montague variant o f Cooper and Parsons (1976) . The use o f subscripted pronouns rather than 5's and x ' s follows Nontague's practice, but that d i s t i n c t i o n I s also immaterial here.)",
                "cite_spans": [
                    {
                        "start": 250,
                        "end": 275,
                        "text": "Cooper and Parsons (1976)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "The semantics o f variable binding i s w e l l studied i n logic; a p a r t i c u l a r l y clear and b r i e f account' can be found i n Quine (1970) . The c r u c i a l p o i n t here i s t h a t the semanti cs i n v b l ves consideration o f a whole range o f possible values f o r the v a r iables, not the determination o f any single value o r referent. Equally crucial i s t h a t the i n t e rpretation o f (2) involves an open sentence w i t h two occurrences o f t h e free variable he i n the position o f the antecedent noud~ri!:, the other corresponding t o the surf ace pronoun .",
                "cite_spans": [
                    {
                        "start": 144,
                        "end": 150,
                        "text": "(1970)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "Using these clear cases, we can discover strong syntactic constraints on the occurrence o f bound variable pronouns. With few exceptions, i t appears t h a t bound varfables must be i n construction wS t h theibr anteceaents ( t h e observation i s made bv Evans (1977); the riotion \"in construction with\" comes from Klima (1964) : a constituent A i s i n construction w i t h a tonstituent B i f and only i f A i s dominated by the f i r s t branching node ,which dominates B. The term c-command i s a more recent a1 ternative name f o r the same notion .) Thus the following do n o t p e m i t a bound variable reading:",
                "cite_spans": [
                    {
                        "start": 317,
                        "end": 329,
                        "text": "Klima (1964)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "(4) (a) Every man walked out. He slammed the door.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "(b) John loves every womanTand he hopes t o date her soon.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "(c) I f no student cheats on the exam, w i l l pass the course.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "By contrast, the bound variable reading i s permitted i n cases l i k e (1 ) and (2) above, i n which the pronoun i s i n construction w i t h i t s antecedent.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "The clearest cases o f what I am ca1 l i n g pragmatic uses of pronouns are cases where a pronoun i s used w i t h no l i n g u i s t i c antecedent a t all; as i n ( 5 ) , o r where the antecedent occurs i n an e a r l i e r sentence of a discourse, as i n (6). I believe t h a t there are no absolute rules governi n g the choice o f referent f o r pragmatic uses o f pronouns, but t h a t there are d i scoverab1.e s t r a t e g i es and p r i n c i p l e $ governing the re1 a t i v e li ke1 i hood o r o r preference among choices. The other p a r t i c ipants i n t h i s panel k n w much more than I do about gz what those p r i n c i p l e s and strategies are ; I hope they would agree t h a t the output of such p r i n c iples i s a probable or expected feferent rather than an absolute referent f o r the pronoun. For example , i n most contexts, the probable referent o f the he i n (6) I s E l l i o t ; but one can e a s i l j enough Z a g i n e a context where speaker and hearer are most interested i n f i g u r i n g out where Max i s , and being unable t o reach E l l i o t i s a good clue t o Max's being i n Boston: then hemay be intended and understood as r e f e r r i n g t o Max. What matters most seems t o be the sallence and relevance o f a p a r t i c u l a r individual, and I see no reason t o draw any theoretical 1 ine between cases where t h a t salience comes from the l i n g u i s t i c context as opposed t o the' non-1 ingui s t i c context. yhere I do want t o draw a sharp l i n e i s between the bound variable use and the pragmatic use of pronouns. The bound variable use i s best described a t the level o f syntactic form and semant i c i n t e r p r e t a t i o n o f single sentences, and the* relevant question i s not what the pronoun refers t o , but what q u a n t i f i e r phrase i s binding it. The pragmatic use i s best described a t the pragmatic level, where the f u l l context o f the sentence i n use i s considtired; on the syntactic level, these pronouns are r e a l l y no d i f f e r e n t from prol per names, and a t the semantic level , they can be viewed as f r e e variables o r as dummy names.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "The basic distjnction.",
                "sec_num": "1."
            },
            {
                "text": "I have begun w i t h the clearest examples o f the distinction; i f a11 use3 o f pronouns f e l l unamb i guously i n t o these two categories, I could stop here. A l I the r e s t would be a matter o f improving the description o f the syntactic cons t r a i n t s on bound variable anaphora and unravell i n g the processing mechanisms t h a t we use t o determine the referents of the pragmatic uses of pronouns. But the clear cases do not provide a set o f necessary and s u f f i c i e n t conditions f o r t e l l i n g the two kinds o f pronouns apart, A l l we can conclude so f a r by way o f conditions i s the following:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "(i ) A pronoun can function as a bound v a n abl e only i f i t i s i n t h e same sentence w i t s antecedent.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "-g ( i i ) Any pronoun can be used pragmatically.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "I f these are the o n l y conditions, we would expect many occurrences o f pronouns t o be ambiguous as t o which use they have, and indeed many are. The pronouns i n (1) and (2) are ambiguous i n t h i s way qnd the sentences have sharply d i f f e r e n t i n t e rpretat-ipns i n the two cases. But now consider a sektence l i k e (7):",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "(7) The prosecutor believed t h a t he would win the case.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "This ebample can be analyzed e i t h e r way; i f the pronoun1 i s analyzed as a bound variable, the sentencd i s interpreted as i n (7a) , and i f the pronoun i s treated pragmatically , we can repre-",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "sent i t ag i n (7b).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "(7a) {The prosecutor: he ) believed that heo would win the cask3 ( 7 b ) The prosecutor believed that he5 would wfn the case.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "On the pragmati c pronoun readtng , the free varl abl e he will be irrterpreted as some salient individual & e n 1 ned by th$ context ; and one 1 i kel y chol ce will,,be the prosecutor. T h i s looks a t f i r s t as i f we are predicting an ambiguity where there i s none.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "And this Ps not just an isolated example, since the same situation wil7 arise whenever we have an antecedent noun phrase that picks out a particular indjvldual. B u t i t turns out that there i s striking evidence that this i s a real structural ambiguity, and not just an a r t i f a c t of the analysis. I belleve that Keenan (1971) was the first t o point this out; Sag (1976) and Wi114ami (1977) discuss such cases extensively. The evldence comes from verb phrasg deleti on. and involves examples 1 i ke the following:",
                "cite_spans": [
                    {
                        "start": 369,
                        "end": 379,
                        "text": "Sag (1976)",
                        "ref_id": null
                    },
                    {
                        "start": 384,
                        "end": 399,
                        "text": "Wi114ami (1977)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "(8) The prosecutor believed that he would win the case, and so did the defense attorney.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "The missing verb phrase can be understood i n just two ways, corresponding t o the two structures we have posi ted for the f i r s t clause On each reading, sentence (8) predicates the same property of the defense attorney as i t predicates of the prosecutor: either the property of being an x such that x believed that x would win the case Tthe bound variable readina, or the property of being an x such t h a t x believes t h a t $ (the prosecutor) would in the case (the pragmatic pronoun reading) .g Thus the examples of so-called \"sloppy identity\" (Ross 1967) of pronouns are really examples of s t r i c t semantic identity of predicates. Thi s important general S zati on can be captured only by recognizing that apparently unambiguous sentences like (7) are i n fact structurally ambiguous.",
                "cite_spans": [
                    {
                        "start": 166,
                        "end": 169,
                        "text": "(8)",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "Cases w i t h proper names as antecedents t o pronouns work just like (7) and (8), the unified treatment of a1 1 noun phrases, i ncl udi ng proper names, as quantifier phrases proposed by Mon tague (1 973)",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "i s an inportant aid in permitting the treatment of pronouns advocated here.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "Another major source af pronoun ambiguity i s t h e systematic ambiguity of most plural noun phrases as betvleen a \"group\" reading and an \"individual\" readihg , aS i n (9).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "(9) Three men lifted the piano.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "Wheh the plural pronoun they I S used as -ragmat i c pr~noun , i t always refers to a group; b u t when kt 1's used as a bound variable, i t may be e l ther a variable over individuals o r a variable over groups, Thus we get two bound vari able readings plus a pragmatic pronoun reading f o r (10).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "(10) The Democrats voted for their, wives. Each of these sentences is ambiguous between a bound variable use and a pragmatic use of the pronoun; and sentence (13a) permits either the individual -level bound vari able reading (each of the two believed she could escape) o r the grouplevel reading (both believed that both could escape). However, (l3b) on the bound variable reading must be a group-level pronoun, because the antecedent is i n a partitive constructton, which requires a groupdenoting noun phrase. A f u l l e r discussion o f plural noun phrases and bound variable prdnouns can be found i n Bennett (7974), although Bennett does not specifically discuss the pragmatic uses of pronouns. No new principles of pronoun interpretation are needed f o r these cases beyond the important observation that they can function semanti cally as an individual-level pronoun, that is, just l i k e a singular pronoun. The complexities of these examples result simply f ron the joint interact1 on of several indtvi dual l y simple phenomena: bound vari able vs. pragmatic uses of pronouns, individual vs. group readings of plurals, and the possd b i l i t y of either a whole noun phrase o r a subpart o f i t serving as antecedent f o r a pronoun.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "The examples distussed so f a r are sumnarized and extended i n Table I below. The column headed \"Pragmati c Pronoun\" should be understood as f 01 1 ows : the given pronoun can be interpreted as referring. to an individual or group determinable on the basis o f the interfiretation of the given \"antecedentM as the relevant 1 ingui st1 c context. Thus, for example, while every man does not refer t o the group of all man, i t can promote that group into salience,_ as can no man and no men.",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 64,
                        "end": 71,
                        "text": "Table I",
                        "ref_id": "TABREF1"
                    }
                ],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "(14) No students came t o the party. They thought they weren ' t i nvi ted.6J when John in*=) tion. However, beither 1 nor anyone else t h a t (b) John spoke t o Mary when he walked in. I know of ever succeeded i n s t a t i n q a version o f But such a view requi res that semantic interpreta-the r u l e which covered a1 1 o f these cases without t i o n operate on surface structure, since the appli-generating c l e a r l y unacceptable resul t s as we1 1. cation o f the r u l e changes the meanly whenever Recent arguments by Terry Parsons ( ersonal comnunthe repeated noun phrase i s anything other than a ication) , Robin Cooper (forthcoming , Gareth Evans Given t h a t pragmatic pronouns must be generated d i r e c t l y anyway because o f cases where there i s no 1 ingui s t i G antecedent, there i s then no work l e f t f o r such a transformation t o do; i t s i m p l i f i e s neither the syntax nor the semantics. Hence i t has been abandoned by l i n g u i s t s o f j u s t about every theoreti cal persuasion. On the defensible assumption t h a t a .donkey should be analyxed here as an existenttal q u a n t i f i e r phrase having narrower scope than the eyeryi e:hi s i t cannot be analyzed as a bound variab e -Partee 1975a)-But i t also does not r e f e r t o any s p e c i f i c donkey, and so does n o t appear t o be functioning as a pragmatic pronoun. Geach suggests t h a t a sentence l i k e (17) be analyzed i n terns o f (18) :",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "Ju4a",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "(18) Every man who owns a donkey beats the donkey he owns.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "reference of a noun phrase 1 i ke the clock t n (20).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "(20) Did you wind the clock?",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "That the missing ingredient i s pragmatic has long been recognized ; Cooper (forthcoming ) proposes a mechanism t h a t brings i n ppagmatics i n a simple (22) the x such t h a t team (x) and R (Jdhn, x).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "What i s comnon t o these analyses of pragmatic pronouns, definite descriptions, and genitive constructions i s the use o f semantic free variables that are pragmatically assigned parti cul ar va3 ues .",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "Introducing the free variables a1 lows a complete specifikation of the form of the interpretation to be given f o r each sentence a t the semantic level, while providing an appropriate division of 1 abor be tween seman ti cs and pragmati c s . 3~ the determination of the content.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "Cooper's proposal for the dmkey and paycheck sentences i s that pronouns can be analyzed not only as-free variables, b u t a1 twnatively as expressions composed of more than one free variable, u t i lizing f r e e property or relation variables much as i h the examples just discussed. The logical fomalism I s cmplex, b u t I will give it f o r completeness and then tby t o paraphrase i t less formally. A singular pronoun (he, -she, or it) may have any translation of the following form:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "(23) 3 x b y [ C p d (Y)? Y = x l A K (XI],",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "where n i s a proper y-denoting expression containing only 'free variables wd parentheses.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "What t h i s says is t h a t e . 9 It may be interpreted as (the pmperty s e t of ) the unique individual",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "x which has property n . For the paychetk,example, an appropriate n w i 11 be R (u), where R i s a free relation variable and u i s a free individual vari abl e that w i 1 1 be bound by the second ocsurrence of the man. The second clause af (19) will theti say \"the man u such that u gives the x such t h a t R (x,u) to U ' S mistress.,'' The pragmatical ly a propriate R will be \"being the paycheck of\". d e computational complex1 ty of the analys is justified, I believe, by the f a c t that only very s a l i e n t relations permit the klnd of pronoun use evidenced by the paycheck example.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "Cooper' s analysl s of the donkey sentences\"ses the identical device; for details see Cooper (forthcoming )",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "The conclusion of t h i s section i s that there are no pronouns of laziness ; the cases which seemed t o requi r$ them can be handled by an extension of the notion of pragmatic pronouns. The extension i s somewhat complex, b u t (a) i t makes use of the same kind o f property and relation variables t h a t are needed f o r an account OT definite noun pnrases and genitlve constructions, and (b) the examples i t is needed for are intuitively complex and i nf requent i n occurrence.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "S t r u c t u r a l l y ambiqwus pronouns.",
                "sec_num": "2."
            },
            {
                "text": "There! are many problems of pronouns and reference ",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Conclusion.",
                "sec_num": "4."
            }
        ],
        "back_matter": [
            {
                "text": "1. There are apparent except1 ons t o even t h i s weak a statement, but I believe they are best understood as i nvol v i ng e l 1 i p t i cal sentences. Consider the f 01 low1 ng exampl e (from Davi d Kapl an, personal cormuni cation) :A: Could a woman become chai man o f the PhJ 1os ophy Depaptment? B: Yes, i f she's qualified. The she i n the second sentence i s not a pra matic p r o n x ; but I t h i n k i t i s best treated as 8 ound by an unexpressed antecedent within' the second sentence, which 5 s not as i t stands a complete sentence, rather than as bourid' by an antecedent i n the previous sentence 3. On the pragmatic pronouh reading, the pronoun he can o f course r e f e r t a someone other than the prosecutor; i n t h a t case the missing verb phrase w i l l a l m s be understood as invof ving reference to the same t h i r d person.4. There i s s t i l l an individual/group ambiguity f o r the subject i n t h i s case, but i t does not a f f e c t the interpretation o f the pronoun, so I w~l l ignore it.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Footnotes",
                "sec_num": null
            },
            {
                "text": ". For s i m p l i c i t y I am ignonng the d i a l e c t t h a t allows the w i t h a singular antecedent; i n that dialect T' 12b) i s as ambiguous as (13).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "5",
                "sec_num": null
            },
            {
                "text": "Not every occurrence o f a quantifier phrase with no has t h i s effect, as the following example from k n s (1977) shows: (I) *John owns no sheep and Harry vaccinates them, The r o l e o f non-1 ingui s t i c inference i n interpreting pragmatic pronouns can be seerl f r o m the following l i n g u i s t i c a l l y similar examples.( i i ) John owns no sheep because Amherst taxes them. (~i i ) John now owns no sheep because Harry poi soned them.I n ( i i ) , them seems t o be generl e shee rather than any group o f sheepi i n (ill) t a most Taus ",
                "cite_spans": [],
                "ref_spans": [
                    {
                        "start": 549,
                        "end": 553,
                        "text": "Taus",
                        "ref_id": null
                    }
                ],
                "eq_spans": [],
                "section": ".",
                "sec_num": "6"
            }
        ],
        "bib_entries": {},
        "ref_entries": {
            "FIGREF0": {
                "type_str": "figure",
                "num": null,
                "text": "(On walking i n t o a room) Why i s he [pointing] here? (6) I couldn't reach E l l i o t l a s t night. -He i s probably i n Boston. These are cases where the pronoun i s being used t o r e f e r t o a p a r t i cular individual , and the determi nation o f whi ch i ndi v i dual the i n tecded r e f erent ii requires making use o f the l i n g u i s t i c and nondl ingui s t i c context. I g n~r i ng some compl i cated cases t h a t I w i l l discuss later, we may say t h a t a t the level o f purely l i n g u i s t i c description, such pronouns function 19 ke free variables which are not bound a t a l l a t the semantic level. A sentence containing one expresses a determinate proposition only re1 ati,ve t o a p a r t i c u l a r choice o f value f o r the varisble, much as a sentence cantaining the word = expresses a determinate proposi t i o n only re1 a t i ve t o a p a r t i cul ar time o f evaluation. Such choices depend on the context of use o f the sentence, whlch i s why I c a l l t h i s a pragmatic use o f pronouns.",
                "uris": null
            },
            "FIGREF1": {
                "type_str": "figure",
                "num": null,
                "text": "12) (a) One of the prisoners believed that she could escape. (b) One of the prisoners believed that they could escape-5 (13) Two of the prisoneh aelieved that they could escape. (a) Two of the prisoners believed that cou1 d escape. I (b) Two of t h e prisoners. be1 ieved t h a t they could escape.",
                "uris": null
            },
            "FIGREF2": {
                "type_str": "figure",
                "num": null,
                "text": "Are there \"pronouns o'f laziness\"? Both t r a d i t i o n a l gramnar books and early transformational accounts sukh asLees and Klima (1963)   suggest a treatment o f pronouns d i f f e r e n t f rom e i t h e r o f the two I have described. This i s the view that a pronoun i s a substitute f o r a linguisti cal l y identical noun phrase; (l5b) would on t h i s vfew be derived f r o m (15a).Thus the i t i s viewed as standing f o r a descript i o n reco=able i n a complex way from the i n t t i a l p a r t o f the sentence. Geach may o r may n~t have called t h i s an example o f a \"pronoun o f laziness\"; the term i s his, but i t has been used by him and others i n a variety o f w v s . What a l l uses o f the term have i n common i s the idea t h a t some pronouns should be analyzed neither as bound v a r i -abl es nor as d i r e c t l y referenti a1 , but f n terms o f some syntactical 1y def4 nab1 e re1 a t i on t o an antecedent noun phrase. Another example f o r which a \"pronoun o f 1 at1 ness\" treatment has plausi b i l i $ y i s (19), f r o m Karttunen (1 969) : (19) The man who gives h i s paycheck t o h i s wife i s wiser than the man who gives i t t o h i s m i stress. This i t t s also not a ~o u n d variable nor d i r e c t l y r e f e r g t i a l ; i t seems t o be a substitute f o r the expression h i s check. I n both Partee (1970) and Partee n*i ued for the existence o f a syntacti c pronoun-of -1 azi ness rule, i ntenddi t o cover both these examples and those cases of what I am now c a l l i n g pragmatic p,ronouns i n which t h e antecedent i s i t s e l f' a d i r e c t l y r e f e r r i nq exoress i on such as a proper noun o r i d e f i n i t e W d e s b i p -( I 5 ) (a) John 'poke to",
                "uris": null
            },
            "FIGREF3": {
                "type_str": "figure",
                "num": null,
                "text": "proper noun or a d e f i n i t e description. ! (1977) , Emmon Bach (personal comnuni cation), and others have convinced-me t h a t there i s no .way t o (16) (a) ~o h h l o s t a watch and B i l l fourid a watch =3 (b) John l o s t a watcb and B i 11 found it.",
                "uris": null
            },
            "FIGREF4": {
                "type_str": "figure",
                "num": null,
                "text": "there are some cases t h a t look as though they make the notion of \"prqnoun o f laziness\" coherent without reduclng i t t o one which covers only a small subclass of the pragmatic pronouhs and hence does no useful work.khat then can we say about the paycheck sentences and the donkey sentences? Many l i n e s o f attack are being explored currently; one t h a t I fSnd p a r t i c u l a r l y promising i s proposed by Cooper (forthcoming) , who suggests a rather natural extension o f the notion o f pragmatic pronoun t o handle them. Before d e s c r i b i~g h i s proposal, I need t o fill i n some background. might be better handled v i a a syntactic substi-t u t i o n r u l e than by e i t h e r the bound variable Russell's ana ~y s i s o f singular d e f i n i t e descripo r the pragmatic treatment. One class was i n t r o -tions (Russel 1 1909') requires t h a t there be a duced by Geach (1962), who provides examples unique object s a t i s f y i n g the description i n order l i k e (17): for the expression t o denote anythtng , and hence notoriouslv f a i l s t o account f o r the successful (17) Every man who owns a donkey beats it.",
                "uris": null
            },
            "FIGREF5": {
                "type_str": "figure",
                "num": null,
                "text": "way t h a t para1 l e l s the account o f pragmatic pronogns given above (which i s also basically Cooper' lie proposes f o r d e f i n i t e descriptions a semantic interpretation 1 i ke Russel 1 ' s but w i t h the addit i o n o f a free property variable P: the clock then denotes (the property set o f ) the unique individual & such t h a t c l o c k ( g and P(x). -A t the semantic level, P is just a free variable; it is l e f t t o the pragmatic interpretation of the sentence i n context to determine an appropriate choice f o r P. Irl a context where there is no satient d i r t i n y i shing property, the singul ar defini te descripion would indeed be inappropriate o r uninterpetabl e. Cooper' s treatment can be seen as a formalizatior of the informal gloss of the7 (by Ka.ti: and others) as \"contextirally d e f i n~t e A s a second background step toward Cooper's proposal , consider the interpretation of genitive phrasq like tbat in (21). (21) John's team lost again.",
                "uris": null
            },
            "FIGREF6": {
                "type_str": "figure",
                "num": null,
                "text": "t h a t I have not touched on. I have not discussed reflexive pronouns, f i r s t and second person pronouns, pronouns in modal contexts, the procomnon nounone, anaphoric determiners like same, different, o r other, OF any of a host of other topics crucial* t o a fuqler account of %he role of pronouns i n reference. ,In some cases the problem is just lack of space and time, b u t i n other cases there are s t i l l d i f f i c u l t open problems. I hope'-that some of what I have included i s ref atively clnfamili ar anb'potenti a l l y useful for computational 1 anguage processf ng endeavors, and I count on my fellow panelists t o f i l l i n some of the holes I have l e f t .",
                "uris": null
            },
            "TABREF0": {
                "content": "<table><tr><td>Yet another source of structural ambiguity is the fact t h a t noun phrases may have other noun phrases embedded wf thtn them, and a pronoun may have either the whole noun phrase or a subpart as antecedent. Sentences (12a) and (12b) do not have this parti -</td></tr></table>",
                "type_str": "table",
                "num": null,
                "html": null,
                "text": "that group might be the Democrats themselves, b u t might be some other group determined by the context. Agairr the three readlngs lead t o corresponding readings tn sentences w i t h verb-phrase deletion:(1 1 ) The 'Democrats voted far their wives before the Republ i cans di d.I will not enumerate the readings, but i t can be seen that the positing of the three ttridttures for the first clause plus the requi remeht that verb phrase deletjon be interpreted as setnantlc identity of predl cati on makes the correct predi ctions about the possible interpretations of the"
            },
            "TABREF1": {
                "content": "<table/>",
                "type_str": "table",
                "num": null,
                "html": null,
                "text": ""
            }
        }
    }
}