HI

... this is an expanding selection of pics and of some of my shorter pieces of writing ... and other bits and pieces ... in German and mainly English ... and other strange languages ... COME BACK AND CHECK IT OUT ... COMMENTS WELCOME

Saturday, August 21, 2010

A REVIEW OF THE FIRST SPECIAL ISSUE OF BIOLINGUISTICS, Vol 3, No 2 (2009)


A REVIEW OF THE FIRST SPECIAL ISSUE OF BIOLINGUISTICS, Vol 3, No 2 (2009)

Wolfgang B. Sperlich, 2009



The first of the nine articles by Koji Fujita sets the tone, if not agenda, with a bold and programmatic statement that takes the Minimalist Program to a grand conclusion: all you need is MERGE. His defense of the anti-lexicalist position may be controversial but it is plausible enough. That syntax generates the lexicon may be an idea whose time has come.

One would have thought that the articles to follow would elaborate on these bold proposals and provide empirical support. Oddly enough, only a few articles seems to do that. Some seem designed, by varying degrees, to counter Fujita’s minimalist framework (I will address each one separately below). One wonders if is an intentional design to foster some kind of debate or perhaps follow the Popperian idea that science is essentially concerned with falsifying the latest hypotheses. Biolinguistics seems to have engaged in this concept before in Vol.3, namely in the ‘debate’ between Postal and Collins where the former contribution, The Incoherence of Chomsky’s ‘Biolinguistic’ Ontology, flies in the face of any reasonable debate, what with Postal’s main gripe being that Chomsky cannot be bothered to reply to Postal’s outpourings. Attacks on Chomsky is an industry in itself and I fail to see why Biolinuistics as a fine testament to Chomsky and his work in linguistics has to publish such nonsense just to be seen fair and open to scientific debate.

Since the articles - those that seem to counteract the Biolinguitics program and  Fujita in particular -  are nowhere near as malignant as Postal’s article, one can look at them in terms of a benign debate and as such take issue with the arguments presented.

As such we can take issue straightaway with the second paper by Jacqueline van Kampen, somewhat unnecessarily provocatively entitled The Non-Biological Evolution of Grammar: Wh-Question Formation in Germanic,. The author doesn’t actually go as far as the title suggests - how could she? - but grapples with the question of parameter setting to distinguish English and Dutch. Her proposal seems to be that Wh-Question Formation in child language acquisition (English vs. Dutch) as a parameter fixing exercise is not only achieved by language input - how else could it be? - but that the parameter itself is established by language input. Nevertheless she is driven to make vast generalizations:

The learnability approach relativizes Chomsky’s poverty of the stimulus, but affirms his position that language is ‘perfect’ in the sense of being learnable as a cultural construct without the assumption of innate grammar-specific a prioris.   

She does Chomsky no favours at all in claiming that she at least agrees with his idea that language is ‘perfect’, especially as she follows up with equating language as a cultural construct - a really bad idea if only because ‘culture’ seems to be one of the more ‘imperfect’, if not silly, constructs known to mankind. What is interesting in her research is of course the question of how and why parameters seem a key to language typology. Jason Kandybowicz gave a good account in his article Externalization and Emergence: On the Status of Parameters in the Minimalist Program, Biolinguistics Vol.3 No.1, sadly not referred to in van Kampen (nor does she refer to two other Biolinguistics articles on parameters by Thornton & Tesan, and by Uriagereka). Obviously language input cannot create typological parameters, as say between Dutch and English, and then somehow select for one or the other. If language is learnt through input alone it cannot have recourse to UG or any other languages. Van Kampen has discarded many powerful explanations for why this is a most unlikely scenario.

In fact van Kampen’s scenario is much more usefully applicable to Second Language Acquisition (SLA). Assuming that SLA individuals have blocked their access to UG by varying degrees - having fixed their typological parameters via child language acquisition - we can explain the various difficulties second language learners have - with emphasis on ‘learning’. Learners have to reconstruct second language parameters from input and from what their implicit and explicit knowledge is about their native language and language in general.

Van Kampen also might have been wise to consider whether or not her research focus is in fact part of UG - if not, there is no case to answer. This is a point made later in Parrott’s article (Danish Vestigial Case and the Acquisition of Vocabulary in Distributed Morphology), namely that many linguistic phenomena, variables in particular, are outside the narrow syntax:

… general Minimalist perspective (Chomsky 1995, 2000 et seq., see also Hauser et al. 2002), where only the operations of the narrow syntax are genetically endowed; all variation is restricted to ‘lexical’ features and the interfaces between narrow syntax and language-external cognitive and sensory/motor systems.

The third article by Anna R. Kinsella, Gary F. Marcus, Evolution, Perfection, and Theories of Language, also seems to be fully anti-programmatic, and one wonders why. The authors say that

Here it is argued that language is neither perfect nor optimal, and shown how theories of language which place these properties at their core run into both conceptual and empirical problems.

This is of course at odds with various MP proposals that language, as a computational system, is optimally designed. Of course this is quite a baffling assertion and in terms of evolutionary progress one may well ask, as did Kandybowicz (see above), why it is that we have all these different languages with different parameter settings. Wouldn’t an optimal solution have no parameters, i.e. why don’t we have just UG as a fully functioning language? Kinsella and Marcus however focus on various features that purport to demonstrate the general imperfection of language, features such as ‘ambiguity, redundancy, irregularity, movement, locality conditions, and extra-grammatical idioms’. For the question on ‘movement’ they should have referred to Fujita who demonstrates Move as a sub-class of Merge and as such is of optimal design. One can take issue with all of the other features mentioned. For example the question of redundancy: one can quite easily turn the argument around and say that redundancy is an essential element of all complex systems - as is well known in information theory and computer program architecture. It is popularly argued that the human brain consists to 90% of back-up systems (back-up systems are redundant by definition). English 3rd person, present tense, verb agreement is a case in point: in the sentence ‘X sits at the table’, X is at least partially recoverable as ‘he, she’ or ‘it’. One can also use this as an example to show how syntax generates the lexicon. Kinsella and Marcus skate on thicker ice with ‘locality conditions’ as for example the question of reflexive anaphora long distance binding remains a highly contested research item across languages. All in all, this article falls rather flat. Perhaps there is also a hint that the authors fail to make a clear distinction between langue and parole, thus allowing for the quite valid assertion that the functions of language invariably are associated with the ‘imperfection’, if not outright stupidity so widespread amongst the ruling classes.

The fourth article by Hiroki Narita, Full Interpretation of Optimal Labeling, is a technical account of whether bare phrase structure is a good idea or not - siding with the latter conclusion and as such seemingly somewhat anti-MP. On the other hand these is no fundamental disagreement with MP, rather a fine-tuning that favours some level of labeling. I lack the technical expertise to determine whether or not Narita makes a convincing case for labeling but I am impressed with the dense arguments put forth. In fact the article strikes me as a worthwhile technical discussion that may well have implications for future directions in MP.

The next paper by Dennis Ott entitled The Evolution of I-Language: Lexicalization as the Key Evolutionary Novelty, is in stark contradiction to Fujita’s anti-lexicalist stance. While the lexicalist arguments are framed within MP, one would have hoped for a better appreciation of syntax. As such, one can turn Ott’s thesis up-side-down and arrive at a good argument for Fujita. We can do this by a series of quotes from Ott’s paper, his point of departure being:

… Hauser et al (2002) speculating that the I-language (syntax and the lexicon) may indeed be the sole evolutionary novelty that allowed humans to cognitively outplay even their closest evolutionary relatives.

We can only agree, especially as ‘syntax’ is mentioned before the ‘lexicon’. Ott then reverses this order and concentrates on the lexicon alone:

Paul Bloom’s (2000: 242): “Non-humans have no words and a relatively limited mental life; humans have many words and a much richer.mental life. This might be no accident.”

We can only concur if Bloom puts ‘syntax’ before ‘words’. To make amends Ott has ‘syntax’ included in the next quote:

The intricacy of semantic properties of lexical items is enormous (Pustejovsky 1995, Chomsky 2000), and there is no evidence for comparative complexities in animal calls. The same is true with regard to structure: at most, animal calls have linear-sequential structure, but no higherorder hierarchical structure as evidences in human syntax.

Chomsky has always struggled with the lexicon, at times doing his best to ignore it as mere material for insertion, or giving it some credence by postulating some sort of mini-grammar attached to each lexical item. This notion seems to reappear in Chomsky’s latest attempt to come to grips with the lexicon, and Ott seems impressed:

A lexical item (LI)] has a feature that permits it to be merged. Call this the edge feature (EF) of the LI. … The fact that Merge iterates without limit is a property at least of LIs — and optimally, only of LIs, as I will assume.EF articulates the fact that Merge is unbounded, that language is a recursive infinite system of a particular kind. (Chomsky 2008: 139)

Perhaps we should ask Chomsky where this ‘edge feature’ comes from, how it arises. What is certain however is that Chomsky understands the minimalist syntax of Merge as the key generator of sentences. Chomsky has never proposed a theory of the lexicon. Ott however does:

Evidently, if this were true, an evolutionary account of I-language would be significantly simplified, in that syntax itself would follow from lexicalization (assignment of an edge feature).

This seems an extraordinary reversal of the facts. How can a mysterious ‘edge feature’ give rise to an iterative structural system that uses Merge as its primary generator? Since Ott himself uses the metaphor of ‘words being building blocks for syntactic structures’ we can demonstrate by analogy: if squareness is an ‘edge feature’ of bricks one can hardly claim that this feature determines the structure of the building - admittedly it may put some minor constraints on it. Fujita’s innovative and elegant proposal is exactly the other way round: the structure of the building determines the building blocks. As such lexical items acquire their ‘edge features’ from syntax and are then inserted accordingly. If we call lexical categories like ‘verb, noun, preposition’ and what have you, edge features, we have a perfectly good explanation for their origins, namely from syntax.

Following this line of inquiry we can agree with the old fashioned concept of lexicalization whereby structures generated by Merge become lexicalized in some instances. Strangely enough, Ott seems to echo just such a sentiment in his concluding remarks:

Rather, the sudden addition of recursive syntax, paired with a capacity for lexicalization, plausibly led to the explosive emergence of symbolic thought that paved the way for modern human behavior.

Next in line is Jeffrey K. Parrott’s paper entitled Danish Vestigial Case and the Acquisition of Vocabulary in Distributed Morphology. This is an interesting investigation into vestigial versus transparent case features as found in various Germanic languages. The former allow for mismatches while the latter do not. If case was a feature of narrow syntax we would not expect this to happen for even Vestigial Case applications. The solution is of course that the Vestigial Case phenomena are part of the morpho-phonological component - as associated with the lexicon - and as such outside narrow syntax, as stated in general:

… general Minimalist perspective (Chomsky 1995, 2000 et seq., see also Hauser et al. 2002), where only the operations of the narrow syntax are genetically endowed; all variation is restricted to ‘lexical’ features and the interfaces between narrow syntax and language-external cognitive and sensory/motor systems.

Interestingly a logical consequence might be that case - vestigial or transparent - is outside narrow syntax. Parrott puts it like this:

If Case features are checked in the narrow syntax, then Case is endowed by UG and available to the child without any need for learning from environmental input. If that were the case, it is hard to see why anything like the transparency constraint would be operative. Even a small set of pronoun allomorphs ought to be sufficient to signal the correct mappings of phonological features to Case features. But if case features are only assigned/realized postsyntactically, say by morphological rules that refer to syntactic structures (McFadden 2004), then these rules too must be learned on the sole basis of environmental input and would thus be subject to transparency.

Case as such becomes a lexical edge feature, and if we follow Ott and ultimately Fujita, we may assume that since the lexicon arises from syntax, there is always the possibility that in this process certain syntax features get transferred to and/or mapped onto the lexicon.

It is perhaps not surprising that the next offering with the funny/provocative title Sex and Syntax: Subjacency Revisited, by Ljiljana Progovac, also contains some funny example sentences:

(32) This is a book that the more you read, the less you understand.

(34) He is a linguist — (as) you know. Parataxis

(35) He is a linguist, and you know it. Coordination

(36) You know that he is a linguist. Subordination

The author also makes some strong claims, like:

Despite the sustained effort of about forty years to analyze Subjacency, to
date, there has been no principled account, with the most recent attempts
faring not much better than the initial proposals.

An anonymous reviewer upbraids her for it but she is not deterred and goes on the counter attack. At some level one has to admire such self-belief but it carries with it the danger of falling flat on the face - from a considerable height. While this is not to suggest that ‘the more one reads her paper the less one understands it’ nor that ‘she is a linguist, and you know it’ should be attributed to Progovac, her style of writing does invite a certain amount of levity. In any case now that she has our attention, what shall we make of her arguments? Her basic assertion is that language evolved in three steps (see also examples 34-36 above):

(A) Parataxis/Adjunction stage, with no hierarchical structure,
where prosody/suprasegmentals provide the only glue for
merger (Jackendoff 1999, 2002).
(B) Proto-coordination stage, where, in addition to prosody, the
conjunction provides all-purpose segmental glue to hold the
utterance together.
(C) Specific functional category stage, where, in addition to
prosody, specific functional categories provide specialized
syntactic glue for clause cohesion, including tense elements and
subordinators/complementizers. It is in this stage that Move
seems to become available.

This scheme contradicts, as she says, the one-step explanations favoured by the likes of Chomsky, as in:

… the influential language evolution hypothesis, according to which Merge (which subsumes Move) was the only evolutionary breakthrough for syntax: Once it emerged, it was able to apply freely and recursively (Hauser et al. 2002, Chomsky 2005, Fitch et al. 2005).

Her arguments are centered on the proposal that Subjacency/islands are proto-language constructs with no possibility of Move. She posits that Move becomes available only at stage three (or C as above). This puts the erstwhile theory in its head which said that Move was freely available from the start and was perhaps only later constrained by islands. If Merge subsumes Move, I am not sure if there is a logical possibility that Move emerges, as it were, at later stages of language evolution. Once we get Merge, don’t we get the whole package? Progovac’s detailed points about the interpretation of subjacency are quite convincing at times but she really exaggerates her speculative mode when she seriously suggests that the finer points of syntax were acquired via ‘sexual selection’:

This communicative advantage is concrete enough that it could have been targeted by natural or sexual selection.

Surely she must have realized that her source quote from Lightfoot (1991) was in jest:

Subjacency has many virtues, but … it could not have increased the chances of having fruitful sex.

If not, Progovac may have to put up with the frivolous suggestion that current sexual selection does not favour linguists.

The second-to-last article, also by Ljiljana Progovac - with co-author John L. Locke - entitled The Urge to Merge:Ritual Insult and the Evolution of Syntax, must have been selected to enlarge on the entertainment value engendered by her Sex and Syntax paper. The authors’ penchant for foxy titles is amusing enough were it not for the content that must fall into the category of - as Chomsky might say - not impossible but highly unlikely.

I have always liked Labov’s (1972) treatment of ritual insults in his Language in the Inner City and it may be worth to remind readers that his point was to demonstrate that ritual insults - as exemplified by black English vernacular - are not some sort of primitive verbal behaviour but instead exhibit complex syntax. The present authors - who have overlooked Labov - seem to hark back to more primitive times when they introduce mouthwatering prospects even in the abstract:

But, is there evidence that such verbal duels, and sexual selection in general, played any role in the evolution of specific principles of language, syntax in particular? In this paper, concrete linguistic data and analysis will be presented which indeed point to that conclusion. The prospect will be examined that an intermediate form of ‘proto-syntax’, involving ‘proto-Merge’, evolved in a context of ritual insult.

Picture the scenario: two cave men ritually insulting each other in front of a cave woman who will then select the winner for sexual procreation - the winner being the one who stumped the opponent with an insult so clever that the other guy was left speechless. No doubt HE was the one who invented ‘proto-Merge’ and had his linguistically modified genes passed on. Never mind that modern-day ritual insults seem to play no more part in ritual courtship - rather being a device for membership of a vernacular culture peer group (as also pointed out by Labov). In any case one would have thought that emerging human courtship rituals would favour linguistic devices that demonstrate compliment (of oneself) rather than denigration (of the other), if only for the simple reason of economy. It is much more efficient to sell oneself by accentuating the positive than waste a lot of energy and time to denigrate a potentially large number of different competitors - this is the first rule of advertising. While there are many examples of the puffed up male demonstrating his superior qualities in terms of physical and mental make-up, there is of course always a final recourse to simple violence. The guy with more muscle wins the lady’s heart. - not that she has a choice if the choice comes down to this level. Another downside to this whole argument must be the role of the female: did SHE not have any input into proto-Syntax? As a cynic one may well congratulate the authors that, conceptually at least, they have hit the nail on the head - a stupidly aggressive male having shaped our linguistic, if not cognitive, evolutionary progress to end up today at the ‘age of stupid’ - in line with the title of a documentary film that seems popular amongst certain sections of intelligent people. Luckily Ljiljana Progovac - with co-author John L. Locke - admit that there could be more to this type of evolution:

It is important to keep in mind that we only claim that ritual insult in the form of compounding was one of the factors contributing to the consolidation of Merge; we are certainly not claiming that it was the only factor. As pointed out by a reviewer, the emergence of (proto-)Merge would have brought about a host of other communicative advantages.

In any case the authors single out VN compounds as ‘living fossils’, suggesting that an item like ‘daredevil’ somehow links back to ‘ancient’ proto-Merge operations. Since Serbian seems to have an abundance of VN ritual insult items, the Serbian language seems to be elevated to equally ‘ancient’ status. I am somewhat concerned about the author’s loose play with historical linguistics, connecting highly speculative terms like ‘proto-syntax’ or ‘proto-Merge’ with synchronic language data. Historical linguistics does conduct serious research on the basis of proto-languages, for example the backtracking from extant Polynesian languages to proto-Polynesian and further back to proto-Oceanic and finally to proto-Austronesian. I say ‘finally’ because historical linguistics is not concerned with speculative time-depths beyond the major language families of the world. The tools of historical linguistics are comparative phonology, morphology, syntax and semantics and it is difficult enough to arrive at a first-level proto-language, let alone reach time-depths of say five to ten thousand years. Furthermore if we accept the tenets of glottochronology - using lexicostatistics - whereby a language tends to change internally every 200 years to be mutually unintelligible, then how can we possibly claim that extant VN compounds declaiming ritual insults are potential ‘fossils’ of a time depth ranging from a 100,000 years to a few million?

Progovac with her fine knowledge of Serbian might have done herself a better service had she at least found supporting items from proto-South Slavic, or better still from proto-Slavic. Even then she could not make the link between any proto-language and the misused ‘proto-Merge’. This is not to say that evolutionary linguistics is not worthy of study - indeed it can be a fascinating thing to do - as attested by some of the papers presented in this special issue. Since this field of inquiry must remain highly speculative, one should test any given hypothesis first on the grounds of common sense in order to avoid highly fanciful accounts that are all possible but are also highly unlikely. Progovac is obviously a very good writer - but is it good science? Chomsky says that science operates like the drunk who looks for his lost keys under lamp post - because that’s where the light is (Barsky 1997:95). By analogy one would wish that Progovac gets a bit closer to the light and subsequently make great discoveries.

Last but not least we have The Third Factor in Phonology by Bridget Samuels. This contribution returns to the full support of the Biolinguistics program:

… I explore the idea advanced in many recent Minimalist writings that phonology is an ‘ancillary’ module, and that phonological systems are “doing the best they can to satisfy the problem they face: To map to the [Sensorimotor system] interface syntactic objects generated by computations that are ‘welldesigned’ to satisfy [Conceptual-Intentional system] conditions” but unsuited to communicative purposes (Chomsky 2008: 136).
                                                
No doubt considered by many (e.g. Pinker and Jackendoff) a radical proposal but logical enough when one follows the likes of Chomsky and Fujita. Samuels provides convincing arguments that a wide range of animals are capable of phonology. Her references are spot on; for example its is interesting to learn that ‘we know that rhesus monkeys are sensitive to pitch classes — they, like us, treat a melody which is transposed by one or two octaves to be more similar to the original than one which is transposed by a different interval (Wright et al. 2000)’. That some humans should have chosen tone as a major phonological feature, thus becomes somewhat less mysterious. The Dr Doolittle scenario also becomes less of a joke: given our shared capacities to vocalize, we can arrive at a rudimentary communication with our pet animals in particular. There is a funny Francis of Assisi sketch where he figures out that he cannot talk to the birds by imitating their singing but by hopping around as birds do when on land - well, according to the present theory he should have stuck with singing. Samuels is however cautious when she says that ‘the organization of bird song is particularly clear, though it is not obvious exactly whether/how analogies to human language should be made’.

Samuels also does well to reference a number of studies that make the point that human language is not necessarily predicated on our erstwhile ability to vocalize but may instead have emerged from ‘action’:

Perhaps, then, the precursors of linguistic syntax should be sought in primatemanual abilities rather than in their vocal skills” (Byrne 2007: 12; emphasis his). I concur that manual routines provide an interesting source of comparanda for the syntax of human language, broadly construed (i.e., including the syntax of phonology). Fujita (2007) has suggested along these lines the possibility that Merge evolved from an ‘action grammar’ of the type which would underlie apes’ foraging routines.

Samuels in her conclusion points out that the popular perception of language = speech is in fact based on a false premise:

Perhaps we should amend the ‘speech is special’ hypothesis: speech is special (to us), in just the same way that conspecific properties throughout the animal kingdom often are; but there is nothing special about the way human speech is externalized or perceived in and of itself.

This approach also buries all the silly arguments about Chimsky vs. Chomsky since it is now clearly established that syntax (Merge) was the evolutionary leap that distinguishes humans from animals.

Perhaps above conclusion should equate with one for this whole review. Nevertheless since the editors of Biolinguistics, Kleanthes K. Grohmann & Cedric Boeckx, ‘encourage everyone, as with anything else published in Biolinguistics, to submit commentary and criticism’, I’ll add a few more words. Was it a good idea to have a special issue? Sure, it was, notwithstanding some of the papers that were off the mark. It is good to read contributions from emerging talents - especially as the selected papers came from a conference on Biolinguistics, Acquisition and Language Evolution (BALE), organized by post-graduate students. Since I have omitted to mention the Guest Editorial: Introduction to BALE 2008 by Nanna Haug Hilton, allow me to make amends by way of this postscript and quote her raison-d’être:

The research interests and backgrounds of us doctoral students on the committee varied across a number of linguistic sub-disciplines. It soon became apparent, therefore, that finding a topic for the conference could prove problematic. After some debate, we concluded that instead of addressing a subject that a few of the students specialized in, the conference theme should be one that unites different linguistic disciplines. The topic that emerged deals with a question that, in our opinion, is at the core of all linguistic research: What are the biological underpinnings of language, and what is the interaction between the innate knowledge of linguistic structure with the language input to which we are exposed?

My only gripe with her question in italics is that the second part of the question seems strangely redundant: since the ‘language input’ we are ‘exposed to’ arises from the ‘innate knowledge of linguistic structure’ in the first place, we are faced with a feedback loop rather than an ‘interaction’ between seemingly independent phenomena. I doubt this was the author’s intention as her first part of the question addresses the crux of the matter.


NOTE: all references but two are from the original articles and not listed below




REFERENCES

Barsky, Robert F. 1997. Noam Chomsky: A Life of Dissent. Cambridge MA: The MIT
  Press.

Labov, William. 1972. Language in the Inner City: Studies in the Black English
 Vernacular. Philadelphia: University of Pennsylvania Press. 

No comments:

Post a Comment