This book presents a new theory of grammatical categories - the Universal Spine Hypothesis - and reinforces generative notions of Universal Grammar while accommodating insights from linguistic typology.
AUTHOR: Ludlow, Peter TITLE: The Philosophy of Generative Linguistics PUBLISHER: Oxford University Press YEAR: 2011
Dennis Ott, University of Groningen
SUMMARY ''The Philosophy of Generative Linguistics'' by Peter Ludlow (henceforth, L) is a collection of L's views on philosophical topics related to the theory of Generative Grammar. As stated in the Acknowledgments, various parts of the book have previously appeared in presentations and papers by L and collaborators.
Chapter 1 (''Linguistic Preliminaries'') presents some basics of linguistic theory. L sketches the transition from Standard Theory to Extended Standard Theory and subsequent developments (so-called ''Government & Binding Theory'' and the Principles & Parameters framework). Topics addressed include X-bar theory, movement rules and islands, generative semantics, quantifier raising, the architecture of the ''Lectures on Government & Binding'' model (Chomsky 1981), and the syntax-semantics interface.
Chapter 2 (''The Ontology of Generative Linguistics'') takes as its starting point Chomsky's (1986) distinction between E-language and I-language. L adds to this the notion of ''psy-language'' (language as a psychological state of an individual), ''a thesis about the language faculty that is part of our cognitive psychology, but which we do not suppose to supervene exclusively on intracranial facts'' (p. 48), in turn based on his view that psychological states can be widely or narrowly individuated. L then goes on to suggest that the term ''knowledge'' in the context of language (''knowledge of language/linguistic rules'') ought to be replaced with a notion of HAVING linguistic knowledge. He discusses the relation between linguistic data (such as judgments), linguistic phenomena and linguistic theory, concluding that the theory attempts to explain phenomena, for which data provide evidence.
Chapter 3 (''Data, Intuitions, Judgments'') makes the case ''that linguistic intuitions are best described as linguistic judgments and [that] they are reliable sources of data.'' L argues that the linguistic intuitions relevant to linguistic theory are instances of judgment, elicited in experiment and reflecting to some degree the knowledge encoded in the internal grammar; he then goes on to argue that such judgments are by and large reliable, addressing various objections that have been raised against the value of linguistic judgments. The chapter contains further discussion of scientific methodology, such as a rejection of naïve falsificationism, and the limited explanatory scope of internalist linguistic inquiry.
Chapter 4 (''A Role for Normative Rule Governance?'') proposes that normativity should be taken seriously by theoretical linguists, at least to the extent that the notion enters into individual rule-following. L suggests that rule-following by an individual is essentially due to normative guidance by the rules of grammar, and suggests a definition of normativity that meets the needs of such an approach.
Chapter 5 (''Worries about Rules and Representations'') addresses various concerns raised by philosophers against the idea of constructing an explanatory theory of I-language. L discusses Quinean indeterminacy arguments (the facts alone do not allow us to determine WHICH of the possible grammars describing the facts is the one that is instantiated in a speaker) and Kripkean worries about rule-following. As for the former, L appears to follow Chomsky in assuming that broader issues concerning theoretical adequacy do allow the theorist to choose among weakly equivalent grammars; as for the latter, L's position appears to be that these worries may require a more externalist view of grammar (his ''wide psy-language'').
Chapter 6 (''Referential Semantics for Narrow psy-Languages'') probes the prospects of referential semantics within a theory of I-language, now construed as ''narrow psy-language.'' L recaps the arguments presented in Ludlow 2003 and the response in Chomsky 2003, and responds to Chomsky's response. L argues for ''expressivist semantics'' as an alternative to truth-conditional semantics, which he claims is compatible with his conception of psy-language.
Chapter 7 (''Best Theory Criteria and Methodological Minimalism'') discusses the role of simplicity and formal rigor in theory construction. L argues that the only real motivation for either desideratum is that it reduces the ''cognitive labor'' on the part of the scientist familiarizing herself with a given theory.
The appendix consists of an interview with Noam Chomsky conducted by L, also available online (http://www.youtube.com/watch?v=CHS1NraVsAc&).
EVALUATION Contrary to what the fairly general title of the book might be taken to suggest, L's book is not an introductory textbook; rather, it presents L's personal views on various philosophical issues related to linguistic theory. As such, the book is certainly relevant to linguists interested in the bigger questions behind the research program they are engaged in, although they are likely to see their excitement about the book curbed by various shortcomings.
The overall organization of the book is rather loose, in a way that leaves the reader puzzled at times. It remains unclear, for instance, what relation the contents of the first chapter bear to the subsequent chapters, and the choice of topics (such as islands, generative semantics, X-bar theory, etc.) in this first chapter appears rather random. Moreover, at various points L's exposition misses the mark quite a bit: e.g., his discussion of X-bar theory fails to mention one of its central properties (endocentricity), designed to rule out exocentric rewrite rules; examples of quantifier raising, including tree representations, are presented without ever explicitly stating what interpretive result is gained from covert transformations. Such omissions, while trivial to professional linguists, are bound to confuse linguistically naïve readers, who this chapter appears to be intended for. (Likewise the section titled ''The Minimalist Program,'' which, curiously, is mostly made up of a very brief discussion of formal semantics à la Heim and Kratzer 1998.)
I also felt uneasy about what seem to me to be conceptual errors and confusions contained in the book. For instance, L states in the introduction that the first chapter reviews first ''Government and Binding Theory'' (roughly, the theory developed in Chomsky 1981, 1986) and then proceeds to ''the Principles and Parameters framework, which grew out of Government-Binding Theory,'' followed by ''some remarks on the Minimalist Program.'' This is a rather misleading description: P&P is a research hypothesis guiding theory construction (one possible theory being that sketched in Chomsky 1981), and so is the Minimalist Program (which takes over P&P's basic conceptual assumptions and adds some further considerations into the picture). Again, such inadequacies in the presentation may mislead the uninitiated reader quite severely.
On the other hand, from the point of view of a professional linguist, the book elaborates on a number of rather obvious points that don't seem to warrant the space devoted to them. For instance, L explains that
''it seems implausible to suppose that the structures and forms being posited for linguistic objects (including PRO and trace, or, for that matter, word boundaries) can be found in the intrinsic physical properties of either written or spoken tokens alone; there is no interesting sense in which the relevant properties are found in physical properties of the acoustical signals or ink markings on the page. It seems equally implausible to think that these properties inhere in linguistic social relations -- no tacit convention is going to yield the principles of binding theory, for instance.'' (p. 60)
In chapter 3, we read:
''I would argue that linguistic judgments are not judgments about rules, or even rule compliance […]. They are simply judgments about linguistic facts, and these facts are determined by the linguistic rules.'' (p. 69)
At least for researchers engaged in Generative Grammar acquainted with its founding documents (such as Chomsky 1965), such statements and the discussions surrounding them yield little insight, and one cannot help but feel like L is beating dead horses at various points in the book.
In some cases, moreover, L fails to make explicit his motivation for entering into a discussion in the first place, leaving the reader wondering about the relevance of the argument. The clearest example of this is chapter 4, where L starts out with the proposition that linguistic principles ''normatively guide'' linguistic competence, noting that this is in general not considered a very helpful (or attractive) position. He then continues: ''But perhaps we can still make sense of the idea'' (p. 90), only to enter into a defense of the normative view of linguistic principles, while offering no further motivation for why this would be an interesting view to hold (or an insightful line of reasoning to pursue) after all. The ensuing discussion runs through various formulations of the notion of ''normative rule-following,'' while leaving the reader without any clear indication of why the issue should have any bearing on the theory of Generative Grammar.
My main criticism of L's book, however, is that, for the most part, it loses itself in quibbling about terminology while, at least to my mind, addressing relatively few issues of substance. L's notion of psy-language, for instance, remains rather mysterious (at least its ''wide'' version; ''narrow psy-language'' appears to be equivalent to I-language in Chomsky's usage), and so does its supposed role in theory construction; L's claim in section 2.3 that we should ''distinguish between an agent's grammar [and] the agent's psychogrammar (which is the psychological state in virtue of which I have that grammar)'' does not help in this regard, since no evidence is offered in support of this ontological distinction. (It seems to me that L's discussion would have benefited from taking seriously Chomsky's remarks in the appended interview: ''[Y]ou define a technical notion in the context of an explanatory theory. You don't just define a technical notion out in space'' [p. 175].) Another example is L's criticism of the terms ''know'' and ''cognize'' (as in ''knowing/cognizing a language''), which he proposes should be replaced with the term ''have (linguistic knowledge).'' It remains open of what practical relevance this move is, beyond expressing a terminological preference.
The same criticism applies to L's alternative to referential semantics outlined in chapter 6, termed 'expressivist semantics.' L here responds to rejections of referential semantics as a component of I-language (Hornstein 1984, Chomsky 2000) and attempts to present a non-referential alternative relying on ''attitudes'' (rather than truth values) and ''referential intentions'' (rather than referents), considered ''expressivistically kosher basic elements'' (p. 144). The net effect of this move is that standard semantic types like e, t, <e,t>, etc. are replaced by the types i, A, <i,A>, etc., and compositional operations like predicate modification are rephrased in terms of ''attitude fusion'' and the like. While one may or may not agree with L's intention of ''internalizing'' semantics, the problem is that L's approach is so sketchy that it proved impossible (for the present reviewer, at least) to evaluate whether or not it makes any substantial contribution, or rather presents a mere terminological variant of extant semantic theory. As things stand, one gets the impression that L is largely engaged in attaching new labels to familiar concepts, which is hardly satisfying.
The highlight of the book is chapter 3, which presents a convincing argument for the validity of traditional modes of data collection, relying on linguistic judgments as an important (but by no means exclusive) source. Quoting from Chomsky 1965, L rightly points out that the lack of EXPLANATIONS, rather than a lack of DATA, is the primary problem of linguistic theory, and convincingly argues for the reliability of linguistic judgments (as supported by recent empirical investigations, see e.g. Sprouse et al. 2011). Furthermore, L argues cogently against taking lightly the notion of FALSIFICATION of a theory by apparent counterexamples (a naïve interpretation of which will lead to ''Frankensteinish theories'') and for a non-monotonic notion of theoretical progress, in which 'coverage' of some pieces of data may at times be sacrificed to explanatory depth. L also rightly emphasizes the limited explanatory scope of generative theories of linguistic competence, which are in principle concerned with merely one ''component in a symphony of elements that in concert might explain everything about 'actual use of language''' (p. 87). The considerations presented in this chapter should be useful to beginning linguistics students and as a reminder to professional linguists as well.
Although they can hardly do justice to the broad scope of L's book, the above remarks should make clear that I found many parts of it not very convincing and rather unsatisfying. That is not to say that L's book is uninteresting, however. It tackles a number of highly relevant and contentious conceptual issues, as in chapter 7's discussion of ''methodological minimalism'' and formal rigor, and L's equally contentious positions on these issues provide a fruitful base for further debate. In general, I find L's ambition to elucidate some of the philosophical underpinnings of Generative Grammar honorable, although I'm not convinced that the resulting book advances our understanding in these areas.
REFERENCES Chomsky, N. 1965. Aspects of the Theory of Syntax. Cambridge, MA: MIT Press.
Chomsky, N. 1981. Lectures on Government and Binding. Dordrecht: Foris.
Chomsky, N. 1986. Knowledge of Language. New York: Praeger.
Chomsky, N. 2000. New Horizons in the Study of Language and Mind. Cambridge: Cambridge University Press.
Chomsky, N. 2003. Reply to Ludlow. In Hornstein & Antony 2003, 287-295.
Heim, I. & A. Kratzer. 1998. Semantics in Generative Grammar. Oxford: Blackwell.
Hornstein, N. 1984. Logic as Grammar. Cambridge, MA: MIT Press.
Hornstein, N. & L. Antony (eds.). Chomsky and His Critics. Oxford: Blackwell.
Ludlow, P. 2003. Referential semantics for I-languages? In Hornstein & Antony 2003, 140-161.
Sprouse, J., C. T. Schütze, & D. Almeida. 2011. Assessing the reliability of journal data in syntax: Linguistic Inquiry 2001-2010. Ms., http://ling.auf.net/lingBuzz/001352.
ABOUT THE REVIEWER
ABOUT THE REVIEWER:
Dennis Ott received his Ph.D. in linguistics from Harvard University in
2011 and is currently a postdoctoral researcher in the ERC-supported
project INCOMPLETE PARENTHESIS at the University of Groningen. His research
interests in theoretical syntax include movement and locality, ellipsis,
and clausal complementation; he is also interested in issues concerning
language acquisition and the philosophy of language and linguistics.