You are on page 1of 6
CONTENTS List of Contributors Introduction Thomas G. Bever, Jerrold J. Katz, & D. Terence Langendoen ‘The Fall and Rise of Empiricism Jerrold J. Katz & Thomas G. Rever ‘The Influence of S} Thomas G. Bever un ipeech Performance on Linguistic Structure 65 Finite-State Parsing of Phrase-Structure Languages and the Status of Readjustment Rules in Grammar 89 D. Terence Langendoen A Dynamic Model of the Evolution of Language Ms Thomas G. Bever & D. Terence Langendoen Analogy or Ungrammatical Sequences That Are Utterable and Comprehensible Are the Origins of New Grammars in Language Acquisition and Linguistie Evolution Thomas G. Bever, John M. Carroll, Jr, & Richard Hurtig A Case of Apparent Ungrammaticality D. Terence Langendoen 149 183 Dative Questions: A Study in the Relation of Acceptability to Grammaticality of an English Sentence Type 195 D. Terence Langendoen, Naney Kalish-Landon, & John Dere ‘Acceptable Conclusions from Unacceptable Ambiguity 225 2. Terence Langendoen viii CONTENTS : cated a Not Sad Ove? san Not Uabappy Person Be Calle Pdi Langencoen Thomas ©. Beer ‘The Argument from Lurk Robert M. Harnish Social Differentiation of Language Structure ‘Manfred Bierwisch Logical Form and Implicature Robert M. Harnish ses and Prevoposiion Told J Kats & D, Terence Langendoen interpretation Global Rules and Surface Structure Interpr Jerrold J. Katz Index 239 261 mm 313 393 415 a7 FOF CONTRIBUTORS* Thomas G. Bever (1, 11, 65, 115, 149, 239) Columbia University Mangred Bierwisch (271) Deutsche Akademic der Wissenschaft 2 Berlin John M. Carroll, Jr, (149) Columbia University John Dore (195) Baruch College, City University of New York Robert M. Harnish (261, 313) University of Arizona Richard Hurtig (149) Educational Testing Service, Princeton, New Jersey Jerrold J. Katz (1, 11, 393, 415) Graduate Center, City University of New York Nanoy Kalish-Landon (195) California State University of Sacramento D, Terence Langendoen (1, 89, 115, 183, 195, 225, 239, 393) Graduate Center, City University of New York "*Nembers in parentheses indicate the pages on which the auchocs! contebutions begin, 4h the late 1960s and early 1970s, generative grammar changed from a stable subject with a unified theory, called the “standard theory” and codified in Katz, and Postal’ Integrated Theory of Linguistic Desriptions (196) and Chomsky's Aspects of the Theory of Syntax (1965), tom less than stable one with « number Of different and confieting theories. Not surprisingly, attitudes toward this change in the state of generative grammar vary depending on which of these theories one is committed to. Most generative grammarians have abandoned the the standard theory in favor of one or another ofthe theories thet emerged in te ate 1960snd easy 1970s, and accordingly, they view the change as progress fovard 4 more adequate mode! of grammar, Some have changed theit theory many tines and so have an even stonger sense of progress. Nonetheles, almost all generative grammasians view the overall condition of thei field with some dismay. The proliferation of diverse theories, they can agree, isnot a state of Mois in which the field cam take pri, Moreover the reaction of scholars and e- searchers in related disciplines such as peycholinguistie, anthropological lngus- Ys, and philosophy of language has been unfortunate. These scholars and ‘eseurchers, who olten have to employ grammatical theory in their studios, ‘have been confused about which theory to employ, and there is now a trend {0 ty to avold grammatical theory entirely by resorting to & purely dateorented ‘pproach, In his situation, it ems ceasonable to rave the question whether the change that took place when the standard theory was abandoned really repre- sented progress ioward a more adequate theory of grammar Inthe ease ofeach ofthe new theories of grammar to which generative gram acing tuned, the change from the standard theory was motivated by the dese to handle phenomena within grammar that did not seem to fit naturally into the available formulation of the standard theory. It was thought that the SMandardtheary could not, in principle, zesommodate the new phenomena and ‘i ROLD J. KATE THOMAS 6. BEVER, JERROLD ). [AND D: TERENCE LANGENDOEN ee aRetne Deck phenomena should realy be hand in grammar, and i, Feea eu re somainey gavel phenomena, We tak tht the ssde theory has many theoreti abvarags ofr thy hod popsed ce osideration. Ina sense, however, a reconsideration of this issue is alcea : Sater ny, ‘Ths weeoigrton has bee aking lise uated insti of uanons that ae somingly uveawe co te ues sounding he debate ‘and relevant of these studies, and focuses them on the que emma nreconidertn of he datas ofthe anda beng. rip sanded then contains te flowing pens The sac of Pine mare eeting ep afm is rae et Tena intunpeetatons Spat rasformationd ees wate the undoing “SO Haier fastens fo espe phre maker and thereby rp Ea the sounesingcoiaton nthe exten, ‘Te undiving me ne ‘sentence of a language are generated by a set of context-free noes and oy ocd tenfomatonsl ns that ir exe tens ino Giceealoane stext-free phrase-structure rules. . The super the structures generated by the cont a ined by applying transformational rules, ficial phrae mares ate obtained by applying transformational iss, WS PE enorme py a 2 wk eelea¥y, fom the ot Gry mt ote or org the eas her of axon rama. Chom Shy cent ight nto grammatealdensption was that axon amma inrmopuction sentences cannot be represented adequately in their surface structures, but rather in their deep structures, This was the point of the now classic examples “John is easy to please” and “John is eager to please”; “Flying planes ean be dangerous"; and so on. The crucial feature of the standard theory was its post lation of a level of deep syntactic structure with special properties that give it a unique place in the explanation of grammatical phenomena. These properties, that the entities at this level are the input both to the semantic component and ‘o the syntactic transformations, established the deep structure as the unique level that determines the semantic interpretations of sentences (relative to the semantic rules) and theit phonological interpretation (elative to the global ‘transformations in the syntactic component and the phonological rules). On the standard theory, then, the all-important grammatical relations connecting sound snd meaning are located far beneath the surface structure of sentences. ‘The standard theory was also the optimal embodiment of Chomsky’s ant ‘empiticist and antibehaviorist criticism of taxonomic grammar end structuralist linguistics generally. By making both the semantic and the phonological struc- ‘ures of sentences depend on their deep syntactic structure, which is both lunobservable and too abstract to be inductively inferable from regularities in the surface structure of actual utterances, the standard theory takes the strongest antiempiricist and antibehaviorist position. No other plausible theory could State more clearly that the objective reality of a language consists notin corpora of utterances (actual or possible) or in stimulus-response relations between cavironmental features and verbal behaviot, but rather in a speaker's inter alization of its grammatical rules, that is, in intesnal psychological states, Siniledy, no other plausible theory could state more clearly that the acquisition of grammar consists not in making inductive generalizations from observable distributional regularities but in developing innate schemata (which represent LUnguistic universal) into grammatical rules that embody hypotheses about the unobservable deep structure of the language. Although much work early in the last decade fleshed out the standard theory and refined the formalization of theoretical insights (see for example Rosen- baum, 1967, and Chomsky, 1970) in the latter half of the sixties andl the early seventies the emphasis shifted, and increasingly work in generative geammat has come to be concerned with developing alternatives to the standard theory. ‘As codified in the carly sixties, the standard theory was incomplete on several ‘important points; the philosophical ideas and assumptions it embodied were not fully digested nor generally embraced; the philosophical and logical azgumenta- ‘don Chomsky introduced into linguistics was not well understood; the relations of grammar to the various disciplines that it now touched upon were unclear; and finally, deseribing natural languages using the apparatus of generative gram, mar was difficult, ‘The alternatives to the standard theory fall within either the generative semantics framework or Chomsky’s extended standard theory. For generative semanticists the fundamental source of dissatisfaction was theoretical, while for 4 ‘THOMAS G. BEVER, JERROLD J. KATZ AND D, TERENCE LANGENDOEN Chomsky it was almost entirely empirical. Ultimately, in both cases the issue came down to how well the various theories could handle the grammatical facts, to a blend of theoretical and empirical issues Generative semantics was the first alternative to appear. Its practitioners believed that the arguments supporting standard theory over taxonomic grammar had not been cated fer enough and, if extended to their logical conclusion, ‘would eliminate even the level of deep syntactic structure, leaving only = highly hxtract and homogeneous mapping of semantic structures onto surface struc- tures (see MeCavley, 1968; Lakoff and Ross, 1967; Lakoff, 1971). Initially, the fanve was between generative semantics and interpretive semantics, which was that part of the standard theory that embodied the partioular thesis thet gram= nara are gytactiealy based (Gesivations ac initiated by the rules that generate linderiying phase markers) and thatthe meaning of a sentence is obtained as an Interpretation of its syntactic structure, in particular, its underlying phrase mnavkers, ‘The question was whether the theory of semantically based graromars being proposed by generative semantcsts was really new and simpler, or merely 2 notational variant ofthe standard theory itself, as both Katz (1971) and Chom- Sey (1972) argued. But before any clear conclusion was reached, a new set of issues arose with the development of Chomsky’s extended standard theory. The generative seruanticists then saw the extended standard theory, which of course Wwas a3 interpretivist as the standard theory, as the “theory to beat.” Their Tssumption was that if Chomsky extended his theory, abandoning the standard dheory.it was pointless to argue directly against the latter a position Chomsky no longer held, Chomsky's prestige in the field at the time made this postion almost fnresistible; and indeed few linguists resisted it. But logically, of course, the fssumption is only a strong as the reasons for the extension ofthe standard theory. If these reasons were insufficlent, then perhaps the change in th theo tetieal scene, too, was misguided. We shall return to this question below. “The main stimulus for Chomsky’s extended standard theory were linguistic phenomena that were ¢iffcult 10 account for within the framework of the Standard theory. These phenomena included quantifier relations, the inter ction of quantifiers and negation, topic-comment relations as in acive/passive pair, focus and presupposition, and certain apparent constraints on the opers tion of transformational rules. It was clalmed that grammatical description Should in principle be able to account for such phenomena, yet the standard theoiy seemed unable to do s0. As Chomsky and a number of his students and colleagues saw it, the problem lay in the standard theory's restriction against femantic rules applying anywhere but at the evel of deep structure. ‘They Srgucd that the recletrant Linguistic facts could be described in grammars that allowed semantic interpretation rules to apply to both underlying end superficial phrase markers. Their contention that certain aspects of semantic interpretation Docur at the love of surface syntactic structure was « denial of the Katz-Postsl 5 peropucrion thes tht synasti ansformaons make no contibuton tothe meaning of tricot 2 Choy 1971; an fackoncoe, 1072) They dae tha theory of gammar sould contaln unless forurface terete ran he seat orponen. Such suds interpretation res, Kate pot is tention (this vole, constitute a major extension ofthe ge of posite grammars, not simply of the range of objects tha undergo murs epe fon within a gamma. Thee sls are ificenty moe pret snee Aeron they can make use ofthe struct of noncontiguous phrase markers. semantly bated amar in order fo cea’ the homogeely of am rail cle Contre sant oppose te extenied stand teary. ‘tht ated ay ving rel components to ts peo Spall” Pos (1972) props « homopnscis framework of eat Simple phase sirctue eae, which define the notion “wellormed tomatic Stata,” and eration consi, which deine the ported clos between sane rpresnatons snd srfacstrstrs, The ue has ero Which iseae inthe power of ganar, that advocated by te pene semantics inder the name “debvaoal constant” or a asvcred ended nda thot under te many “suerte” essa 10 acount forthe phenomena tha bth take Seay tf he enone at th eo eS More recently th ue has been frther compat on both side. There ts gute ements (kare, 1973, ad Labfl, 1973) wh rete a rn tor powell app, “easel cone which operate beyond the context of 2 singe detivton and block cet Uerztins om te ats of oer dortons Tis as il he move sharply ausson ofthe elton betwen nsw inthe homogeneity of aleané he Sze ofthe die of pomble granu: how much ofan ieee in he st of ossble gammae ein b flerted atthe pace of increas i homens? Whar eonierations can letinaey enter into desing ths question? Mos basal is clu thatthe face tat moval sich adel incase nga lvepouer ace not desrbblin tomo le power parts of sme Kn? noma snd hi flowers further comple th iss Charly (1973) and Feng (74) now propos the reve extend stoned tect whch al] semantic interpretation occurs at the level of surface structure. "ss Katz, pols oun is conton to ths volume, he advantage of the eed ex teed standard tory shat takes “bal” ues unpecsary 8 hig inerpretaton osc at the sume level. For this, howe es nessa (0 Crrch the suras srt, so that can epee al ofthe state stee Tey tp epee ton hese more sent referent fn neste cans andthe exten sandadhory do otter the somunon sumption weigher Oat 6 THOMAS G. BEVER, JERROLD 1. KATZ, AND B, TERENCE CANGENDOEN standerd theory was abandoned for good reason; and thet one or the othe: of these two theories is preferaéle to the standard theory. The present volume challenges this assumption. We think that it hes been accepted too uncritically, and that it is based on 2 misconception about the linguistic phenomena motivat ing revisions of the standard theory. We think that the standard theory is priori preferable to any of these extensions by virtue of its being the clearest expres- sion of the original insights of transformational grammar and because it const tutes the narrowest constraints on the class of possible grammars. Thus, we ‘think that linguistics stands to gain much if it can be shown that the best ex- planation of linguistic phenomena does not require extensions of the standard ‘theory. "As we have already remarked, what fostered the development of the altemative theories was the belief that certain linguistic phenomena were (a) grammetical phenomena and (8) could not be handled in grammars in the frame- ‘work of the standard theory. Although there has been much work in the latter trea to demonstrate that linguistic phenomena cannot be handled in the standard theory, this means nothing unless it can be shown that the phenomena in ques- tion are indeed gremmiatical in nature. Forexample, it is meaningless to argue for a pasticuler theory of intelligence on the grounds that that theory can, and other theories of intelligence cannot, handle a certain set of psychological facts if those psychological facts turn out to be about accident proneness rather than in- telligence. Hence, what is necessary is a careful, systematic examination of the claim that the linguistic phenomena that motivate extensions of the standard ‘theory are indeed grammatical rather than extragrammatical facts of enother sort, |As we have indicated above, some investigations of this issue have been con- ducted, with, however, quite different aims and without any connection to each other. What we have done is bring some of the more important of these studies together in this volume as a first step toward a systematic examination 0° the issue of what exe grammatical phenomene. ‘The picture revealed by integrating these independent studies is the follow ing For almost all linguistic phenomena that have provided empirical support for extensions of the standard theory, the assumption that they should be describable within 2 theory of grammatical competence is, at least, highly ‘questionable. Recall that the early Chomskyan conception of grammars as heo- ries of the internalization of a system of rules, and of linguistic theory as the initial competence of the language leamer caused work in the description of Janguages to touch on a number of disciplines in a way that it had not ecrlier. ‘Because the relation of grammar to these disciplines, in particular psychclogy, philosophy, sociology, and logic, was so new, it was as yet unclear. This meant the boundaries of grammar and linguistic theory themselves were uncleat, too. If, for example, we cannot say where the psychology of sentence peroep- tion begins, we are also unclear about where the grammar of sentence structure ‘ends, At the same time, linguists working in grammar were successful in de- 7 INTRODUCTION shits a or of rate ota tht a ete in aa The new taoratona ods ta an ine of econ poled pootiy tna oe ‘in previous theories; and grammarians were busy expl. th of this "ela ie sleet Ino few ett ate. Dt, won ty ce ange think that one o¢ another of the restctions imposed’ 94 the aes oe ‘epenoe vena On son, nate goat weeds ae Heron wi il gana Sn Chon eer es of intelligibility in multiply center-embedded sentences is a matter of fore mance factor reflecting the intation on immedate memory rales than mater of gamma. But onc ese estan! phenome teens re si eer they neve fominy amma eed mee fry no gua ee we da nyo oe hy Se pane phenome Iolo hi © sae ee ie ae fo cin tht ack pean fe en mesos yen Bsus nd nctingy > 2" pone suit apret Re eaa Pit wt wnat oe yoga, ase cde linet sr knde of thous Ie Glow, phiosopy,shetons, socilogy—tat account Toe ey many of to bye sme wey that Chmsi accouned fo ow of itelipiiy sean Snbedded sentenss. They donot only show bow sich phonons cabo Sule y ts tli se uo rumar urbe meg oft bonis enn pean he Segue aa Toy tae wept vos hypotess sou the wage a re tnd a abou i etn fo nelgboning daspnes ese een rove sound, much of the alleged evidence agains! roan Soper ofthe rss teats daar nana Or adn seat Ee newman do ese sob that an pin sc ar etna ee es fos Sol petit ole Gly ees ses Som ht hone sch een cna! any saeco ony gr mm Ts a “Ie fens of ees oman on Une Sects” by BoA Dyce ads Sebo, nf Langan y Bees ageing ey Sees Hg; A Cw of Apu ganna" angesace Se a {cs by Laren, Kah Canon snd Boe aan sos ae "erson Be Called a Not Sad One?” by Lan a he i Pee Bc 9 Tatendou od Bows Ofc tee tue seta ae threes to wintry vi ‘per paceptin sn proton an augue ae 8 THOMAS G. BEVER, JERROLD}. KATZ [AND D, TERENCE LANGENDOEN ‘phenomena, where univocal treatment in the grammar would necessitate abat doning the standaré theory. Katz's “Global Rules and Sarface Structure Interpretation,” points out the rmuanitude of the extension embodied in the extended standard theory and sug gests that a major segment of the evidence for this theory is irclevant to the issue, because it is stylistic or rhetorical rather than grammatical in mature “Pragmatics and Presupposition” by Katz and Langendoen, and Harnish’s papers “etle Argument from Lurk” and “Logical Form and Implicature” provide strong grounds for thinking that various extensions of the standard theory in generative semantics were mistakenly based on pragmatic sather than grammatical phe- nomena. “Social Differentiation of Language Structure” by Bierwisch is com cemed with the proper treatment of sociolinguistic phenomena outside gran: mar. “FiniteState Parsing of PhraseStructure Languages and the Status of Readjustment Rules in Grammar” by Langendoen further articulates the sten- dard theory; it provides a formalism for the notion of a syntactic readjustment rule and settles a previously open question sbout the boundary between competence and performance Finally, two general papers address directly the inadequacy of generstiv? semantics, “Acceptable Conclusions from Unacceptable Ambiguity” by ‘Langendoen points out the problems inherent in a current version of the theory ‘of transderivational constraints and shows that certain of the phenomena in ‘question can be handled within the standard theory. “The Fell and*Riso of Empiticism” by Katz. and Bever traces the history of transformational grammar from the work of Zellig Hacris to that of generative semantics, and demonstrates ‘hat ono wing of generative semantics involves a return to the empiriciom of taxonomic linguistics in the 1930s end 1940s. REFERENCES Bever, T.G. 1970, “The Cognitive Basis for Linguistic Structures.” In Cogn- tion and the Development of Language, ed. J. R. Hayes. Wiley. Chomsky,N. 1965. Aspects of the Theory of Syntax. W.LLT. Press 1970, “Remarks on Nominalizations.” In Readings in English Transforms tional Grammar, R. Jacobs and P, Rovendaum, eds. Ginn-Blisdel 1971, “Deep Structure, Surface Structure and Semantic Interpretation. a ‘Semantics, D. Steinberg and L, Jakobovits, eds. Cambridge University Press. 1972, “Some Empirical Issues in the Theory of Transformational Gram- mar.” In Studies on Semantics In @ Generative Grammar. Mouton. 1975. “Questions of Form and Interpretation.” Linguistic Analysis 1 75-108, Fiengo, R, 1974, Semantic Conditions on Surface Structure. Unpublished Ph.D. dissertatation, MILT. 9 inrRopueTion fSkendotl, R, 1972, Semantic Inepremton in i, © Generate Grmar Kat, 11. 1971, “Generative Samant i tne . EE BM, rete Samant Kate, J.J andP Pol 1964. An Isegrate Te ‘ ef doa raed Theory of Linguistic Dsooions ott, 6. 1971, “On Generate 6 1. Sakobonts, es, Cambriage Uiveriy ress > Stmers an 1913. “On Transdervatinal Constraints” jonal Constraints” In Biers In Honor of Hen and Renée Kahane, B. Kachru, ed. Linguistic Research. aa aes laff, @, and 3.8, Rom. 1967, Unive Lingistios Ci. MeCawley, J. 1968. "The Role ofS “nguite Theory, E Bach sou Post, P1972. “The Best Theory Prentice al, Rotenbaum, P..S. 1967. The Grammar of English Predicate Complement |. The Grammar of Engl 7 7 : of Engl ite Complen Linguistic Js Deep Structure Neceseery? Indiana >mantics in a Grammnar.”" In Universals in Harms, eds. Holt, Rinchart and Winston, ” In Goats of Linguistic Theory, S. Peters,

You might also like