In 2015, The MIT Press honoured two of Noam Chomsky's most important works in linguistics, Aspects of the theory of syntax (henceforth, Aspects) and The Minimalist Program (henceforth, MP), with anniversary editions celebrating their original publication 50 and 20 years earlier, respectively. Chomsky contributed new prefaces, which situate the works in their historical context and adumbrate subsequent extensions and revisions in his own work. Although the reissue of these works as an acknowledgment of their enormous influence is commendable, the poor execution on the publisher's part significantly diminishes the practical value of the reprints. Both new editions introduce a number of typographical mishaps, including a bungled quotation on p. 50 of Aspects and the transformation of intonational into international on p. 31 and p. 216 of MP. Moreover, the gratuitous decision not to preserve the pagination of the original editions is more than unfortunate, not least because the reissued chapter 4 of MP cannot be navigated with the help of Zwart's Reference Zwart1998 excellent Rough guide.
There is surely no need to recapitulate the enormous impact Aspects and MP alike had on the field of linguistics. In Chomsky's own words, “Aspects was the first attempt to outline a general framework for the study of I-language” (Aspects, p. x), single-handedly laying the foundations for the unusually fruitful research program of generative grammar (“a grammar [that] purports to be a description of the ideal speaker-hearer's intrinsic competence [and] is, furthermore, perfectly explicit” (Aspects, p. 2)). Many ideas first fully developed in Aspects have, of course, become part and parcel of the program, among them the distinctions between competence and performance, weak and strong generative capacity, and different levels of adequacy. Most of these are lucidly explained in Aspects’ famous first chapter, making it as appropriate a reading for an introductory syntax class today as it was at the time of its publication. As “the real problem” for the research program built on these notions, Aspects identifies “that of developing a hypothesis about initial structure that is sufficiently rich to account for acquisition of language, yet not so rich as to be inconsistent with the known diversity of languages” (p. 60). The tension was resolved, in general terms, by the Principles and Parameters model that forms the point of departure for MP, which adds a focus on “general considerations of conceptual naturalness” deriving from a vision of the language faculty as characterized by “simplicity, economy, symmetry [and] non-redundancy” (MP, p. 1). MP revitalized the field founded by Aspects during a period of explanatory stagnation, proposing a new perspective on the nature of the language faculty in general that informs the analysis of individual phenomena.
Rereading Aspects and MP, one is struck in equal parts by the sharp discontinuities and continuities between these works. In Aspects, Chomsky asks “how detailed and specific […] the innate schema [is] that gradually becomes more explicit and differentiated as the child learns the language” (p. 27) – this being, of course, MP’s guiding research question as well. But unlike Aspects, MP takes as its starting point not the question, but rather a hypothetical answer: what if the innate schema is minimal, and much of the complexity of language arises from the fact that it interfaces with language-external systems pertaining to interpretation and externalization? In practical terms, this program of “approaching UG from below” (Chomsky Reference Chomsky, Sauerland and Gärtner2006) has the therapeutic function of counteracting “the temptation to offer a purported explanation for some phenomenon on the basis of assumptions that are of roughly the order of complexity of what is to be explained” (MP, p. 214). MP and its offspring are thus, in large part, a reaction to the recognition that “the critical problem for grammatical theory today is not a paucity of evidence but rather the inadequacy of current theories of language to account for masses of evidence that are hardly open to serious question” (Aspects, p. 19).
Aspects’ fame derives in no small part from the fact that it introduces a wealth of theoretical innovations that continue to figure prominently in current theorizing (for illustrative examples, see the contributions in Gallego and Ott Reference Gallego and Ott2015). These include, for instance, the introduction of the lexicon as a component separate from the computational system; the reliance on features of lexical items, inspired by phonology; the seeds of X-bar Theory (“subcategorization of V involves only frames that are dominated by the symbol VP” (p. 107)) and the duality of semantic interpretation (based on observations about movement affecting quantifier scope in footnote 9 of chapter 3); and the reduction of grammatical functions to phrase-structural configurations, a precursor of Hale and Keyser's (Reference Hale, Keyser, Hale and Keyser1993) configurational approach to thematic relations. These and other ideas are preserved by MP (albeit in a different form), and other technical proposals – such as the matching analysis of relative clauses proposed in chapter 3 and the idea that case and agreement rules “rewrite terminal symbols in certain restricted ways” (Aspects, p. 188) – have been resurrected in subsequent work by Chomsky and others.
But other, more audacious theoretical ideas introduced in MP were presumably inconceivable in the days of Aspects. Economy conditions on derivations, which have only a weak analogue in Aspects’ evaluation metric (which however evaluates grammars, not derivations), are one example. Chapters 2 to 4 of MP capitalize on the ideas that “derivations must be as economical as possible: there is no superfluous rule application” and “there can be no superfluous symbols in representations” (p. 137) – the principles of Last Resort and (a particular interpretation of) Full Interpretation, respectively. Economy principles are taken to be “specific to the language faculty” (p. 147). In later works, Chomsky speculates that principles of this sort might not be language-specific at all, but instead reflections of efficiency of computation and, ultimately, nature's striving for optimality (the “Third Factor” in language design of Chomsky Reference Chomsky2005).
In yet other cases, MP radically revises central assumptions of Aspects. A particularly striking case is Aspects’ dismissal of the proposal that rewrite rules create sets rather than strings, which “can lead to a study of grammatical relations that is independent of order, this being a phenomenon that belongs to surface structure” (p. 133). Chomsky rejects this approach in favour of concatenation, since “no proponent of a set-system has given indication of how the abstract underlying unordered structures are converted into actual strings with surface structures” (ibid), concluding that there is “no reason to consider the set-system, for the time being, as a possible theory of grammatical structure” (p. 134). Precisely such a “set-system” is, of course, later developed in chapter 4 of MP, coupled with the assumption that “ordering is part of the phonological component” (p. 307ff.). In a partial concession to Diderot, whose proposal that “the order of words is determined by factors independent of language” is dismissed as a “naïve view of language structure” early on in Aspects (p. 6), more recent work by Chomsky views bidirectional mapping relations between syntactic representations and sensorimotor instructions as an altogether ancillary adaptation to sequential externalization/perception.
Similarly, while Aspects embraces phrase-structure rules as a “natural mechanism for generating phrase-markers” (p. 70), already the “pre-minimalist” first chapter of MP (co-authored with Howard Lasnik) identifies them as “a dubious device […], recapitulating information that must be presented, ineliminably, in the lexicon” (p. 22), largely rendered obsolete by X-bar Theory and lexical information. Chapter 4 of MP eliminates much of this residue, introducing in its stead the operation Merge and rules of projection that allow for a restatement of X-bar-theoretic notions in relational terms. Chomsky's more recent works go yet further in arguing that projection rules should be abandoned entirely in favour of an algorithm that detects labels (Chomsky Reference Chomsky2013), a development foreshadowed in MP’s allusion to the possibility of labels being independently “determined” (p. 224).
A leading theme recurring throughout Aspects and MP is simplification of the theory. It is this conceptual desideratum, not empirical considerations, that leads Chomsky in Aspects to separate the base component from the lexicon, so that “grammatical properties of formatives […] need not be represented in the rewriting rules” (p. 93); they are instead specified in lexical entries, resulting in a considerable simplification of the base component. Simplicity considerations are likewise behind Aspects’ rejection of the “generalized transformations” of Chomsky Reference Chomsky1975 in favour of a recursive base component generating “generalized phrase-markers” with clausal embedding, to which (singulary) movement and deletion transformations apply in a cyclic fashion, thereby reducing the grammar to “a base and a linear sequence of singulary transformations” (p. 144). Chapter 3 of MP reverses this picture, reintroducing generalized transformations in the form of a binary substitution operation subsuming Move-α (recast later, in chapter 4, as Move-F, which displaces features). Displacement is construed as the case of substitution where α “is a phrase marker within the targeted phrase marker […] itself” (p. 174) – a reduction that would later lead Chomsky (Reference Chomsky and Belletti2004) to eliminate displacement entirely in favour of a single Merge operation that includes the “internal” case. Only with this later explication would the copy theory of movement, introduced in MP’s chapter 3 as the suggestion that “the trace left behind is a copy of the moved element, deleted by a principle of the PF component” (p. 185), ascend to the status of a true null hypothesis.
These developments notwithstanding, the discussion in MP is still firmly grounded in the assumption that displacement is a departure from minimalist design principles: “chains are not introduced by selection from the lexicon or by Merge” (p. 291). Hence they necessitate a separate operation and indices that distinguish copies from accidental repetitions, “a departure from the inclusiveness condition, but one that seems indispensable” (p. 208). In lieu of a clear rationale for such an operation, Chomsky suggests “that the [displacement] property should be reduced to morphology-driven movement” (p. 204), based on the assumption that uninterpretable inflectional features must be eliminated for the resulting representation to satisfy Full Interpretation. In my personal judgment, this suggestion – which, as Chomsky notes, “tends to eliminate the possibility of optionality in derivation” (MP, p. 133) – spawned the most harmful aberration in the field's recent history, giving rise to a proliferation of proposals postulating ad hoc featural triggers of movement that, in almost all cases, amount to little more than a restatement of the fact that displacement happens (see Richards Reference Richards2016, p. 1). Chomsky's proposal thus inadvertently fueled the very temptation to restate observations in technical terms which he sought to subdue.
The reduction of displacement to Merge in Chomsky Reference Chomsky and Belletti2004 (and earlier works such as Bobaljik Reference Bobaljik, Pensalfini and Ura1995) renders “justification” of displacement in terms of stipulated triggers obsolete, paving the way for a maximally simple Merge operation, applying freely (Collins Reference Collins, Bauke and Blümel2017). Various other technical proposals advanced in MP fell prey to subsequent minimalist scrutiny as well. One example is the numeration, introduced as a device to determine a “reference set” for derivations to render transderivational economy constraints computationally feasible. But when it is stated later that “output conditions enter into determination of the numeration itself” (MP, p. 270) such that an element of the numeration is licensed only “if it has an effect at the interface” (MP, p. 338), it becomes clear that the numeration is nothing other than a new level of representation introduced through the back door, contrary to declared minimalist ambitions.
Some other details of implementation appear to simply have been lost in subsequent technical discussion. For instance, in chapter 4 of MP, Chomsky defines Merge as an operation that “takes a pair of syntactic objects (SOi,SOj) and replaces them by a new combined syntactic object SOij” (p. 208, my emphasis), thus “eliminating” (p. 223) the original syntactic objects from the workspace. Merge, on this view, reduces the workspace, except when some new element from the lexicon is introduced. Much subsequent work tacitly adopted a different view, according to which terms related by Merge remain in the workspace, enabling operations such as Parallel Merge (Citko Reference Citko2005). Technical details of this sort have been insufficiently addressed since the publication of MP, despite the fact that they bear directly on fundamental questions of syntactic computation (see Chomsky et al. Reference Chomsky, Gallego and Ottto appear for some discussion).
An often overlooked conceptual discontinuity between Aspects and MP concerns the notion of grammaticality (the property of being constructed by the grammar). Aspects advocates the development of a “theory of ‘degree of grammaticalness’” (p. 161), despite Chomsky's recognition that “the notion ‘grammaticalness’ cannot be related to ‘interpretability’ […] in any simple way” (p. 162) and an admission that “it is clear that various diverse factors determine how and whether a sentence can be interpreted” (p. 163). Footnote 2 of Aspects’ chapter 4 proposes that “the grammar directly generates the language consisting of just the sentences that do not deviate at all”, whereas it “derivatively generates all other strings, with [structural descriptions that] will indicate the manner and degree of deviance of the derivatively generated sentences.” This general perspective is largely abandoned in MP, where grammaticality (now called convergence) survives only as a theory-internal notion that is not in any obvious way related to perceived well-formedness: “[a] convergent derivation may produce utter gibberish […]. Linguistic expressions may be ‘deviant’ along all sorts of incommensurable dimensions, and we have no notion of ‘well-formed sentence’ […],” and the only empirical criterion to be satisfied by the theory is that “[e]xpressions have the interpretations assigned to them by the performance systems in which the language is embedded: period.” (p. 178). For further discussion of these issues, see Ott Reference Ott2017.
In view of their ubiquity as background references, it is easy to lose sight of the fact that both Aspects and MP stand out as highly unusual publications in the field. Aspects’ ambition to outline a general framework for generative grammar is matched in scope only by MP 40 years later. The general methodology of MP (as well as later publications of Chomsky's) separates it sharply from the vast majority of works in the field as it developed subsequent to the publication of Aspects. Rather than aiming for maximal empirical coverage in any particular domain, Chomsky homes in on specific phenomena (such as expletive constructions, which dominate much of the discussion of chapter 4 of MP) that he takes to be particularly revealing with regard to the nature of the language faculty, often with significant idealizations of the database. In MP’s original preface, Chomsky expresses his “personal feeling […] that much more substantial idealization is required if we hope to understand the properties of the language faculty” and bemoans the fact that “misunderstandings and confusion engendered even by limited idealization are so pervasive that it may not be useful to pursue the matter today”, adding that idealization “is a misleading term for the only reasonable way to approach a grasp of reality” (p. 6). Even a casual glance at the majority of current publications – many of them bearing titles adorned by the vacuous add-on “a minimalist approach” – reveals the sharp discrepancy between this view and the prevailing attitude. It is perhaps the single greatest merit of these reissues that they can serve as a reminder of Chomsky's unmatched callousness in carving up the inextricable mess of phenomena in a way that might offer “a grasp of reality”.
