agrammaticality in performance

13
Agrammaticality in Performance Author(s): Carlos Otero Source: Linguistic Inquiry, Vol. 4, No. 4 (Autumn, 1973), pp. 551-562 Published by: The MIT Press Stable URL: http://www.jstor.org/stable/4177800 . Accessed: 15/06/2014 13:10 Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp . JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. . The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to Linguistic Inquiry. http://www.jstor.org This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PM All use subject to JSTOR Terms and Conditions

Upload: carlos-otero

Post on 20-Jan-2017

214 views

Category:

Documents


2 download

TRANSCRIPT

Agrammaticality in PerformanceAuthor(s): Carlos OteroSource: Linguistic Inquiry, Vol. 4, No. 4 (Autumn, 1973), pp. 551-562Published by: The MIT PressStable URL: http://www.jstor.org/stable/4177800 .

Accessed: 15/06/2014 13:10

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

The MIT Press is collaborating with JSTOR to digitize, preserve and extend access to Linguistic Inquiry.

http://www.jstor.org

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

SQUIBS AND DISCUSSION

meaning at all. While it is not the same as ambiguity (formally, at least) perhaps it does not contrast neatly with ambiguity. Vagueness may be an interesting issue in its own right.

References

Catlin, J. C. and J. Catlin (I972) "Intentionality: A Source of Ambiguity in English?," Linguistic Inquiry III,

504-508. Lakoff, G. (I970) "A Note on Ambiguity and Vagueness,"

Linguistic Inquiry I, 357-359.

AGRAMMATICALITY IN For reasons which need not detain us now, in my recent PERFORMANCE squib (Otero I972) I argued only what seemed the least

Carlos Otero, UCLA controversial part of my case, namely, that strings such as SE alquilan apartamentos, common in ordinary speech, "cannot be directly generated by the grammar of Spanish." However, a more interesting claim is that they cannot be generated at all, either "directly" or "derivatively" (Chomsky I965, IV, fn. 2), because they are in fact "agram- matical" (to coin a new term by analogy to amorallimmoral).1 What is agrammatical is neither grammatical nor "un- grammatical" (taking the latter term in the sense of 'partially grammatical'); it is simply outside the scope of the grammar.

An optimal grammar must establish a sharp division between two disjoint classes: the class G of grammatical sentences that it generates (either directly or derivatively) and the class G of agrammatical strings (nonsentences) that it does not generate at all (cf. fn. Io). But "grammaticalness is only one of the many factors that interact to determine accept- ability" (Chomsky I965, ii). "It is clear that the set of

1 The replacement of the word utterances of my original title by the term sentences disambiguated the word ungrammatical in a sense contrary to the one intended. In the expression ungrammatical sentences, ungram- matical can only mean 'partially grammatical' ('semigrammatical'), as when ungrammatical strings is equated with semisentences (Katz, I964, 400)

or when the feature * which assigns "strangeness" in Chomsky (1971,

Examples (42)-(45)), is said to be "ungrammatical", i.e. partially grammatical. But when, as in Katz's very next paragraph, the view that "a grammar is a formal system whose rules permit the derivation of every grammatical string in the language while, at the same time, insuring that no ungrammatical string can be derived in the system" is attributed to Chomsky, ungrammatical is equivalent to my agrammatical. Since an utterance (a performance unit) can be correlated with the corresponding sentence (if there is a sentential counterpart), it seems legitimate enough to speak of "grammatical", "ungrammatical", or "agrammatical" utterances (cf. Chomsky 1955, passim; 196I, 187).

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

552 SQUIBS AND DISCUSSION

grammatical sentences cannot be identified with the linguist's corpus of observed sentences. Not only are there many (in fact, infinitely many) non-observed grammatical sentences, but, in addition, certain sentences of the corpus may be ruled out as ungrammatical [= agrammatical], e.g. as slips of the tongue" (Chomsky I955, Sect. 23) or, pre- sumably, as slips of the mind.

Important observations regarding the unacceptability of grammatical sentences because of limitations on perform- ance have been made in Chomsky (I965) (cf. Chomsky and Halle I968, 372). But as far as I know no one has yet con- sidered the opposite case, that is, the possibility that utterances commonly accepted as performance correlates of grammatical sentences of a language L might not be generable at all by the grammar of L. Since crucial evidence with a bearing on this question is not easy to obtain, any empirical information which might contribute to improve our understanding of the important distinction competence/ performance is likely to be of value. In this light, the Romance phenomena discussed below (as a follow-up to the aforementioned squib)2 might be of interest to any linguist.

Consider the following Spanish sentences:

(i) a. Las noticias se difundieron. 'The news spread.'

b. Se difundieron las noticias (no la verdad). 'The news spread (not the truth).'

c $SE } difundio las noticias, pero d. Radio Nacional no se difundieron.

'(PROj broadcast the news, but it [lit. they] RN fdidn't spread.'

Ignoring the expression inside the parentheses, (ib) is derived from (ia) by a rule that moves the subject NP to the end of the sentence. The two sentences are of course cognitively synonymous. But it should be obvious that (ib) is not synonymous with the first part of (i c): the second part

2 Familiarity with the squib is assumed. Care should be taken to correct the following: Example (I4) should read se los alquilan (rather than SE ... .); the references in (i ga) and (i gb) should read (i i a) and (ii b), respectively; the gloss for (4c) should read are (not were); after (4c), immediately before We note. . ., insert:

As further evidence, consider the following paradigms:

(4) d. (yo) (*me) vivo 'I live' (tu) (*te) vives 'you live' (61) (*se) vive 'he lives'

(*Nl) SE vive 'PRO lives' me atrevo 'I dare' te atreves 'you dare' se atreve 'he dares'

*SE (se) atreve 'PRO dares'

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

SQUIBS AND DISCUSSION

of (ic) is basically the negation of (ia) and yet the contra- position of both parts of (ic) does not yield a contradiction. Corresponding to the difference in meaning, there is a difference in structure. Thus if Topicalization of the direct object NP (which always leaves a pronominal copy behind) is applied to (i c) the new sentence is (2), while there is no sentence (3):3

(2) Las noticias SE las difundio. (3) *Las noticias se las difundieron.

Further evidence that las noticias 'the news' in (I c) and (2) is a direct object NP rather than a subject NP comes from this paradigm:

(4) a. SE difundio (las) noticias (a las tres de la tarde). 'PRO broadcast (the) news (at 3 P.M.).'

b. Se difundieron las noticias * (a las tres de {las} la tarde). 'The news spread (at 3 P.M.).'

because Spanish requires a determiner in a deep subject NP which is not a proper noun (e.g. los pajaros cantan 'birds sing'). 4

A natural way of accounting for these facts is to assume that the verb difundir is entered in the lexicon of Spanish with this specification:

(5) a. difundir is pronominal or causative b. the subject of the pronominal is absolutive

where the pronominal means essentially 'intransitive para- digmatically reflexive', "the term causative is so defined that a verb with the feature [ + causative] appears in the context subject-object, where the object is the subject of the corre- sponding intransitive", and absolutive is the "case" of "the thing acted upon but not modified by the action" (Chomsky

3 Although identical in surface structure to (3), the surface (i) is identical in deep structure, not to (3), but to (ii):

(i) Las noticias "se" las difundieron. (ii) Ellos/ellasi lesj difundieron las noticias.

'They: broadcast the news for them.'

4 The determiner does not always appear in surface structure (e.g. (algunos) casos se han dado en que... 'there have been cases in which . . .'). Notice also that *se difundieron noticias is not the same as SE difundi6 noticias in structure or meaning. As an "acceptable" pseudo- equivalent of the latter, the first variant of (4b) is of course common among some speakers, especially the "educated" ones (this is precisely our topic).

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

554 SQUIBS AND DISCUSSION

1972, 170-172). Thus the lexical entry (5) specifies that difundir can appear in the contexts of (6) in deep structure:

(6) a. NP1 se (cf. (ia-b)) b. NP2 r 1 NP1 (cf (ic))

[+ CausJ

In line with this, the deep structures for (7a) and (7b) are (8a) and (8b), respectively:

(7) a. Se difundieron las noticias. <= Las noticias se difundieron. 'The news spread.'

b. SE difundio (las) noticias. 'PRO broadcast (the) news.'

(8) a. S

NP1 VP

zli / \ las noticias se difundieron

b. S

NP2 VP

[+PRO ] V NP1 + Human

[+ Caus]

difundio las noticias

There are many Spanish verbs which share the speci- fication (5a) with difundir. For example, if we replace absolu- tive in (5b) by dative (the "case" of "the thing or person acted upon or toward"), (5) would also be appropriate for such verbs as desmoronar, derrumbar, disolver, romper:

(g) a. Se est6an resquebrajando los diques. ?z Los diques se estan resquebrajando. 'The dikes are cracking.'

b. SE esta resquebrajando los diques. 'PRO is fracturing the dikes.'

(io) a. Se derrumbaron los edificios. = Los edificios se derrumbaron. 'The buildings collapsed.'

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

SQUIBS AND DISCUSSION

b. SE derrumbo los edificios. 'PRO demolished the buildings.'

( i) a. Se disolvieron las manifestaciones. <= Las manifestaciones se disolvieron. 'The demonstrations broke up.'

b. SE disolvio las manifestaciones. 'PRO broke up the demonstrations.'

(Observe that (9b) embodies Hanoi's denunciation, while (9a) was offered as an alternative explanation, and that the lexical items of the English glosses are not even the same in (g) or (io).)

However, with some verbs (such as those in (I 3) and (I4)) the context (1 2) is, if not mandatory, at least far more natural than (6a):

(12) NP1 se ADV(erb) (I3) a. Se lavan bien las camisas. -= Las camisas se

lavan bien. 'The shirts wash well.'

b. SE lava bien las camisas. 'PRO washes the shirts well.'

(I4) a. Se vendieron bien los apartamentos -= Los apartamentos se vendieron bien. 'The apartments sold well.'

b. SE vendio bien los apartamentos. 'PRO sold the apartments well.'

The difference in meaning between the (a) and the (b) sentences is now greater than in (9)-(I i) because of the different effects of the adverb on the two different struc- tures: for example, (i4a) can only mean that the apart- ments were easy to sell, while (I 4b) can mean that the price was at least fair or that the salesman did a good job. On the other hand, if we replace vendieron 'sold' by vieron 'saw' there is no (a) sentence (taking vieron in its literal sense) :5

(15) a. *Se vieron bien los apartamentos. -= *Los apartamentos se vieron bien.

b. SE vio bien los apartamentos. 'PRO saw the apartments well (thoroughly).'

From the foregoing examples and those in the earlier squib we can conclude that an optimal grammar of Spanish has to incorporate (I6):

(i6) a. In (7)-(I4) the (a) sentences differ from the (b) sentences in structure and, correspond- ingly, in meaning. In particular, the struc- ture [DET N]NP functions as a subject in the

5Keep in mind again that the (a) and (b) sentences are supposed to be nonsynonymous (see fn. 4).

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

556 SQUIBS AND DISCUSSION

(a) sentences and as a direct object in the (b) sentences (cf. (8), (23)-(24)).

b. The (a) sentences are directly generable only if the verb is, like difundir, specified for (5a) and (6); the (b) sentences are directly generable only if the verb is transitive and can take a [+human] subject.6

We now turn our attention to the strings in question. It is well known that for some speakers an utterance such as (I 7b) is often taken as equivaluent to (I 7a):

(I7) a. SE [vendio6 (los) apartamentos. b. tvendieronf 'PRO sold (the) apartments.'

But there is only one topicalized form of (I 7) (that is, (i 8)), and only one pronominal counterpart (namely, (I9)):

(i8) Los apartamentos SE los {vendio * vendieron

(i9) SE los vendio } * vendieron)

There is also no directly or derivatively generated counter- part to (2 ia) = (2ob).

(20) a. fSE vendio (los) b. lLa inmobiliariaf a partamentos.

'PRO/the real estate company sold (the) apartments.'

(2 I) a. La i bilii f vendio6 (los) aparta- b. t*vendieronf mentos.

As (2 ib) shows, ordinarily there is no number "agreement" between the direct object NP and the verb. Rather, an un- specified subject usually takes a verb form unmarked for number:7

6 There is no (b) counterpart in (i) or (ii): (i) a. Las patatas se pudrieron.

'The potatoes rotted.' b. *SE pudrio las patatas.

(ii) a. Las paredes se desplomaron. 'The walls collapsed.'

b. *SE desplom6 las paredes. (cf. (io)) On the other hand, (iiib) has no (a) counterpart:

(iii) a. *Las chabolas se demolieron. b. SE demolio las chabolas.

'PRO demolished the huts.' 7 In other cases, such as (i)

(i) b {Haban} (muchos) apartamentos.

'There were (many) apartments.'

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

SQUIBS AND DISCUSSION

(22) a. {Hay} (muchos) apartamentos.

'There are [lit., there is] (many) apart- ments.'

b. f Llueve \*Llueven 'It rains.'

That the noun apartamentos is interpreted as the head of the direct object NP (cf. (2)-(4)) in all three cases of (23), and not only in the first two, is shown in (24). Observe that the adverb bien blocks both variants of (24a) when the lack of determiner makes it impossible to take the NP as a subject, although it does not block (24b):

(23) a. Los apartamentos fueron vendidos. 'The apartments were sold.'

b. Los apartamentos SE los vendio. ( i= ( 8))

C E vendieron (los) apartamentos. (= (I7))

(2 4) a.S E vendi6 1 (*bien) apartamentos. (cf. (vendieronf (4))

'PRO sold apartments (well).' b. Se vendieron bien los apartamentos. (=

(14a))

The "rule" responsible for the surface difference be- tween (I 7a) and (I 7b) cannot possibly be taken as a rule of an optimal Spanish grammar since, in addition to its idio- syncracy and to its not being general and undisputed (many speakers do not use it at all, others use it only optionally or erratically, some only in writing), the substitution of a plural verb form in (I 7b) has no effect on the semantic interpretation of the utterance. Clearly, the semantic inter- pretation of (I 7b) does not correspond to its surface structure.8 Rather, it corresponds to the surface structure of (I 7a) (cf. (8b)). This is readily seen if we compare the interpretation of (I 7b) with that of (ib), (9a), and the like

(b) is widely "accepted" among some speakers, especially in Spanish America (where not long ago, it should be remembered, a substantial proportion of the speakers-among the immigrants and the Indians- were not native speakers of Spanish; many present-day speakers still are not).

8 This discrepancy between form and meaning is taken advantage of in Monteiro Lobato's humorous short story "O colocador de pro- nomes" reprinted in D. L. Hamilton and N. C. Fahs, eds., Contos do Brasil, New York, Appleton, I944, 193-210. (I am indebted for this reference to Ronald Harmon.)

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

558 SQUIBS AND DISCUSSION

or with contrasts such as (25), which is interpreted in the sense of the English gloss by every speaker of Spanish:

(25) Los apartamentos no se venden (por si mismos); SE los vende. 'The apartments don't sell themselves; PRO sells them.'

Finally, notice that the difference between (I 7b) and (I 7a) has no effect on the gradation of acceptability of the utterance (cf. Chomsky I 97 1, fn. I 9), while expressions such as (2ib) or (26b) do not share the status of (17b).

(2 6) a. Algie salio b ge salieron}

'Someone left.'

The choice of salio in (26) or vendio in (i 8)-(2 i) admits of no "relaxation". Notice, furthermore, that even the string (26b) is "interpretable" as a puzzle.9 A string can be "agrammatical" strictly speaking and yet close enough to grammaticality so that an interpretation can be imposed on it. It is easy to forget that grammaticality is a theoretical term. It defines what a grammar generates. If "only well-formed grammars are acquired" (Chomsky and Halle I968, 384), we have to assume that the grammar internalized by the native speaker can only be well-formed. But it would be as absurd to ask a native speaker whether a given string is in fact generated by his internalized grammar as it would be to ask him to write down all and only the rules of the grammar or to define in some other way the category of admissible derivations. The linguist cannot expect the native speaker to do his job for him. An informant can only be asked "to judge the acceptability of sentences, and similar tasks" (Chomsky I 97 1, fn. 59; cf. Chomsky I 965, 1.4). So it seems quite misleading to use unacceptable and the ambiguous un-

9 It should be kept in mind that "grammaticalness is not attrib- uted to an utterance by virtue of the fact that the utterance is recog- nized to be meaningful" (Chomsky I96I, I84) and that, contrary to what is assumed in Katz (I964), there are "two kinds of nonsense, grammatical nonsense and ungrammatical nonsense" (Chomsky 1955,

28. I). After all, "the native speaker is able to distinguish an utterance in normal English from an utterance such as Carnap's 'Pirots karulized elatically' or from Carroll's jabberwocky, which conform to all rules of English but are made up of items that happen not to be included in the lexicon of the language" (Chomsky and Halle i968, 295). Katz's set of "nonsense strings", then, cannot be taken to be the set of "agram- matical strings" (cf. now the passing reference in Katz 1972, 63n, and chapter 8). Another basic difference between Katz's treatment and Chomsky's is that for Chomsky "the principles that determine how inter- pretations can be imposed on deviant sentences may be universal" ( 955, IV, fn. 2; cf. II, fn. ii).

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

SQUIBS AND DISCUSSION

grammatical as practically interchangeable terms, as is often done even in some of the best analyses.

From this perspective it is only natural to conclude that the strange "twisting" of the surface structure which pro- duces an utterance such as (I 7b) with an interpretation identical to that of (I 7a) can only be a performance foible, no doubt related to the difficulties inherent in producing and interpreting pairs such as (7)-(io). The fact that some speakers "accept" (i7b) as a semantic equivalent of (i7a) tells us more about possible shortcomings of performance than about the "grammaticality" of the structure. There is an unbridgeable gap between a "deviant sentence" in the usual sense'0 and an "agrammatical utterance". It seems clear that an optimal grammar of Spanish cannot generate strings such as (I 7b) either directly or derivatively. Rather, a speaker with a mind which has a far greater scope than does its internalized system of grammatical rules is tamper- ing with his grammar in one way or another in such cases. A linguist who is free from the "leave your language alone" syndrome of our behavioristic yesteryear ought to be able to recognize a human shortcoming for what it is. After all, people, as distinct from automata, do make mistakes. True, "the grammar is one of the essential conditions of speaker

10 I have in mind Chomsky's view, as presented, for example, in Chomsky (I965, II, 2.3.i). There are other views, such as that recently expressed by Bedell in an exceptional paper (I972, iI): ". . . word strings are not classified simply into sentences and non-sentences, but a series of classes of non-sentences result which deviate from grammaticalness to different degrees" (emphasis added). Echoing Katz (I964), Bedell writes that "the ideas in [Chomsky I955] . .. do not clearly distinguish those ungrammatical [i.e. partially grammatical] sentences which can be understood ('semi-sentences' in Katz' terminology) from ungrammatical [i.e. agrammatical] strings of words" (emphasis added). But what Chomsky was trying to set up was "a sharp division between a class G of grammatical sentences and a class G of ungrammatical [i.e. agrammatical] sequences" (1955, 28.1; emphasis added) and, in addition, an ordering of sentences from more to less grammatical. If the set of "partially grammatical strings" is not taken to be a subset of the grammatical strings (rather than a subset of the agrammatical strings), there is no natural way (cf. Harrison I972, 202) of defining it and assigning to each "partially grammatical string" "a structural description that indicates the manner of its deviation from strict well-formedness." Only a de- scriptively adequate grammar can provide structural descriptions which "will indicate the manner and degree of deviance of the derivatively generated sentences" (Chomsky I965, IV, fn. 2; emphasis added), "when some rule is violated in a derivation", "anywhere in the derivation" (Chomsky I972, I32). It is clear, then, that for Chomsky all strings which are "generable" or "derivable" (be it "directly" or "deriva- tively") are sentences. It should also be noted that the analogy between "degree of phonological admissibility" and "degree of grammaticalness" has its limits. The lack of recursiveness of the phonological component does not allow for ordinary innovative creativity. There are no "phono- logical metaphors", for example (cf. Chomsky and Halle I968, 4I6-4I8).

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

560 SQUIBS AND DISCUSSION

strategy", but "the generation of a sentence according to the rules of the grammar must not be confused with actual speech production. This important distinction has an analogy in arithmetic ... How the individual operations are carried out in mental arithmetic depends ... on a psychological calculation strategy which is not part of arithmetic. It is conditioned, among other things, by the limitations of the memory. A multiplication which cannot be carried out mentally can easily be performed when the capacity of the memory is supplemented by pencil and paper, the same rules, which are stored in the brain, operating here too" (Bierwisch I97i, 66).

If it is possible to depart from the rules in something as simple as multiplication, even with the aid of pencil and paper, why not in speech? Once a mistake is made and repeated, it is even easier to mimic it. The Pavlovian reflexes induced by society are real enough even in speech (no matter how innovative speaking is in its essence), especially when appropriately "reinforced" and associated with a higher level of "education", as is the case with utter- ances such as (I 7b). 11 Nonetheless, if we had a complete, optimal grammar of Spanish (the only way to a definition of grammaticality in Spanish), those utterances will no doubt turn out to be a sort of "fabrication" of the grammar user, completely outside the range of sentences generated (directly or derivatively) by his internalized grammar. The flawless "competence" of a computer incorporating an optimal grammar of Spanish would never come up with (I 7b) or any of its numerous analogues, because a mindless computer cannot possibly be misled into constructing (I 7b) somewhat by analogy with sentences like (ia-b) and yet interpreting it as analogous to (i c), that is to say, as synonymous with (I 7a) .12

11 An even better index of "education" and "distribution", rein- forced by a purportedly "literary" or "elegant" mode of writing (quite pedantic for ordinary speakers), is the archaic enclisis of the pronoun in cases such as violo for the plain lo vio 'he saw him/it' of everyday speech. Surprisingly enough, this archaism is still "acceptable" to some present- day speakers of Spanish (as its Italian counterpart is for some speakers of Italian) even in cases in which presumably it would not have been acceptable to the native speakers of a millenium ago (whose speech is apparently being imitated), in spite of the fact that the appropriate sub- system of rules has not been part of the grammar of Spanish for at least half a millenium (cf. Otero I 972) .

12 Long after this squib was submitted for publication, this journal published a reply to my earlier squib by Heles Contreras (I 973). There is much that seems to me questionable in Contreras's squib (as the careful reader of nmy earlier squib or the present one will immediately realize), but only one point needs to be made here. Contreras fails to see that, al- though his example (i) is often used as a paraphrase of his example (i o) by some speakers, there are many such pairs in which the paraphrase

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

SQUIBS AND DISCUSSION

References

Bedell, G. (I972) "On Three Notions in Chomsky's Logical Structure of Linguistic Theory," Critiques of Syntactic Studies I (UCLA Papers in Syntax, No. I), i-i6.

Bierwisch, M. (I97I) Modern Linguistics, Mouton, The Hague.

Chomsky, N. (I 955) The Logical Structure of Linguistic Theory, mimeograph, MIT, Cambridge, Mass.

Chomsky, N. (I96I) "Some Methodological Remarks on Generative Grammar," reprinted in H. B. Allen, ed., Readings in Applied English Linguistics, 2nd ed., Appleton-Century-Crofts, New York, I 964, I 73-I 92.

Chomsky, N. (I965) Aspects of the Theory of Syntax, MIT Press, Cambridge, Mass.

Chomsky, N. (I97I) Conditions on Rules, to be published by Mouton, The Hague.

Chomsky, N. (I 972) Studies on Semantics in Generative Gram- mar, Mouton, The Hague.

Chomsky, N. and M. Halle (I968) The Sound Pattern of English, Harper and Row, New York.

relation does not obtain (cf. his condition (g) before example (9)), as those given in (I 7) and (i 8) of my earlier squib, which I repeat here, in a slightly modified manner, for the reader's convenience:

(I 7) a. Se reunieron los miembros de la junta. 'The members of the junta got together.'

b. SE reuni6 los miembros de la junta. 'PRO brought together the members of the junta.'

(I8) a. Las cosas se arreglan (por si mismas). 'Things fix themselves.'

b. SE arregla las cosas. [=SE las arregla] 'PRO fixes things.' [='PRO fixes them']

Contreras leaves (I 7b) out of his example (9), although (I 7b) is often used and is readily acceptable for many native speakers (including my- self); he does not even mention (I8). However, it is obvious to every native speaker I have consulted (including whole classes of them) that the (a) sentences are not paraphrases of the (b) sentences. Neither are examples (9a), (ioa), (I Ia), (I3a), or (I4a) of the present paper para- phrases of the corresponding (b) sentences. The (a) sentences are not synonymous to the (b) sentences, as the lack of contradiction in (Ic) of the present paper demonstrates. It should also be clear, in particular from (6)-(8) above, that the structure of the (a) sentences (which are intransitive) is quite different from that of the (b) sentences (which are transitive and causative). But a grammar which contains Contreras's Verb Agreement "rule" (i i) will automatically convert the (b) strings into the (a) strings (presumably leaving their structures otherwise in- tact). In other words, Contreras's grammar cannot generate the causa- tive structures formed with numerous "ergative" verbs like arreglar, desmoronar, derrumbar, disolver, difundir, etc. I am sorry to discover that my earlier squib was not clear enough to prevent misunderstandings of this sort. It was precisely the (late) realization that this possibility was not unlikely which led me to prepare the present squib immediately after I read the printed version of the earlier one.

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions

562 SQUIBS AND DISCUSSION

Harrison, B. (1972) Meaning and Structure: An Essay in the Philosophy of Language, Harper and Row, New York.

Katz, J. J. (I 964) "Semi-sentences," in J. A. Fodor and J. J. Katz, eds., The Structure of Language: Readings in the Philosophy oJ Language, Prentice-Hall, Englewood Cliffs, New Jersey, 400-4 I6.

Katz, J. J. (1972) Semantic Theory, Harper and Row, New York.

Otero, C. (1972) "The Development of the Clitics in Hispano-Romance," to appear in the proceedings of the Conference on Diachronic Romance Linguistics held at the University of Illinois, April 2I-22, 1972.

NEANDERTHAL MAN SPEAKS? In a recent article in this journal Philip Lieberman and Karl Reisman, Edmund S. CrXlin (1 97 i) reported on a study reconstructing

University of Pennsylvania the vocal tract of the fossil from La Chapelle-aux-Saintes and pointing out resemblances of its supralaryngeal vocal tract to that of newborn humans and the implications of these resemblances for the range of sounds that the fossil could be assumed capable of producing.

I am not competent to discuss their actual recon- struction or the comp'iter simulation on which judgments were based. But there are certain relations between the data presented and the conclusions drawn that seem to me to require comment.

As a preliminary it should be said that it is not exactly clear just what is being claimed. The impression of a very dramatic conclusion is fostered by some rather coy play with definitions. On the first page it is said that "newborn humans, like nonhuman primates, lack the anatomical mechanism that is necessary to produce articulate speech." This in itself sounds sufficiently striking. But inability "to produce articulate speech" is immediately given a rather special definition: "That is, they cannot produce the range of sounds that characterizes human speech." That is, if I may paraphrase, they cannot produce the full range of sounds that characterize speech as we know it today. This is not quite the same thing.

More important, what are the implications of the La Chapelle-aux-Saintes fossil's vocal repertoire? What popu- lation does he represent? Whether a dead end of evolution or a far out variant of an ongoing population, La Chapelle- aux-Saintes is a freaky fossil. Even Lieberman and Crelin, in spite of the narrow focus of their physical anthropology, imply this. Any truly objective approach to the problem of the evolution of speech would presumably have chosen a fossil more central to the range of variation of contemporary

This content downloaded from 194.29.185.230 on Sun, 15 Jun 2014 13:10:17 PMAll use subject to JSTOR Terms and Conditions