The Strong Minimalist Thesis
Abstract
:1. The Earliest Formulation of a Strong Minimalist Thesis
andA language, in turn, determines an infinite set of linguistic expressions (SDs), each a pair (p, l) drawn from the interface levels (PF, LF), respectively.
which are followed by a final assumption:Conditions on representations—those of binding theory, Case theory, θtheory, and so on—hold only at the interface, and are motivated by properties of the interface, perhaps properly understood as modes of interpretation by performance systems.
To get from this final assumption to the formulation of SMT in [1] involves some reasonably straightforward substitutions. Replacing the linguistic expressions with language is to a large extent motivated by the assumption cited above that a language “determines an infinite set of linguistic expressions” (putting aside the question of the relation between language and a language3 plus the fact that language “has been studied intensively and productively for 2500 years, but with no clear answer to the question of what language is” [9] (p. 2)). Next, the terms legibility conditions and interface conditions are virtually synonymous given that the first term is actually interpreted as ‘legibility conditions at the interface(s)’. This synonymy is demonstrated in the discussion in [1] of how the SMT might bear on the notion of “linguistic evidence”, where the SMT is interpreted as saying that “an optimal solution to legibility conditions satisfies all other empirical tests as well”4 (p. 97). The equivalence is expressed in the sentence that follows: “The reformulated thesis replaces the obscure notion of “linguistic evidence” by the meaningful notion: satisfaction of interface conditions.” The statement also sets up an equivalence between solution to legibility conditions and satisfaction of interface conditions, where satisfaction in this context is again virtually synonymous with realization, and thus linking the first formulation of the SMT in [1] to the final assumption that defines the Minimalist Program in [2].The linguistic expressions are the optimal realization of the interface conditions, where “optimality” is determined by the economy conditions of UG.
 “unexplained elements of S_{0}” (the initial state of the language faculty).
 interface conditions (“the principled part of S_{0}”).
 “general properties” (which are external to S_{0}, e.g., computational efficiency).
But here we have moved from an optimal solution to a perfect solution of interface conditions.6 Moreover, this formulation is motivated from a consideration of the function of third factor conditions (e.g., computational efficiency) in language design, thereby linking the categories (b) and (c) of [10] mentioned above and, in effect, ‘approaching UG from below’, which considers question “How little can be attributed to UG while still accounting for the variety of Ilanguages attained, relying on third factor principles?” The above quoted passage also provides an extended formulation of the SMT such that all phenomena of language have a principled account, whereby they are derived “by efficient computation satisfying interface conditions.”An Ilanguage is a computational system that generates infinitely many internal expressions, each of which can be regarded as an array of instructions to the interface systems, sensorimotor (SM) and conceptualintentional (CI). To the extent that third factor conditions function, the language will be efficiently designed to satisfy conditions imposed at the interface—one can imagine more radical theses, to which I will briefly return. We can regard an account of some linguistic phenomena as principled insofar as it derives them by efficient computation satisfying interface conditions. We can, therefore, formulate SMT as the thesis that all phenomena of language have a principled account in this sense, that language is a perfect solution to interface conditions, the conditions it must at least partially satisfy if it is to be usable at all.
Presumably the answers to these new questions will come from outside the limited domain of UG and the narrow consideration of language data that is used to motivate proposals about it (as has been the focus from the earliest work on generative grammar), where, under the SMT, the focus has shifted to questions of efficient computation satisfying interface conditions.Insofar as the third factor can be shown to be operative in the design of FL, explanation can proceed “beyond explanatory adequacy” in the technical sense, raising new questions: not only asking what mechanisms suffice to determine Ilanguage from data available, but why these mechanisms should exist, and whether they are real or just dispensable descriptive technology.(pp. 3–4)
Significantly, there is no mention of interface conditions, nor in the entire paper. Instead, the discussion of the SMT appears to have shifted from interface conditions to the formulation of generative procedures in concert with principles of computational efficiency, now focused on a concept of minimal computation.In the best case, phenomena would be explained by interaction of the simplest computational operation Merge, with its two logically possible subcases. Internal Merge IM (automatically yielding “the copy theory of movement”) and External Merge EM—interacting with general principles of minimal computation MC. The Strong Minimalist thesis SMT articulates this goal.
This explication is then qualified in the following words:The basic principle of language (BP) is that each language yields an infinite array of hierarchically structured expressions, each interpreted at two interfaces, conceptualintentional (CI) and sensorimotor (SM)—the former yielding a “language of thought” (LOT), perhaps the only such LOT; the latter in large part modalityindependent, though there are preferences. The two interfaces provide external conditions that BP must satisfy, subject to crucial qualifications mentioned below. If FL is perfect, then UG should reduce to the simplest possible computational operation satisfying the external conditions, along with principles of minimal computation (MC) that are language independent. The Strong Minimalist Thesis (SMT) proposes that FL is perfect in this sense.
As formulated here, the SMT involves three factors: computational operations, interface conditions, and principles that determine computational efficiency. What follows here will take up each in turn.SMT is not precisely formulated. MC can be interpreted in various ways, though some of its properties are uncontroversial, and reliance on these carries us a long way, as work stimulated by MP has shown.
2. Interface Conditions
The speculation that conditions on sound and meaning “may be further specified by the interaction of grammar and other systems” (presumably performance systems) is a first step towards the view that the systems that interface with Ilanguage might actually constrain their form and function, which has become a central hypothesis of the Minimalist Program.Representations in UP and LF provide the “interface” between linguistic structures and the performance systems. Thus, we assume that all aspects of phonetic form and meaning determined strictly by sentence grammar are indicated in the representation in UP and LF, respectively. The grammar, then, determines a soundmeaning relation; more accurately, it determines a pairing of conditionsonsound and conditionsonmeaning, each of which may be further specified by the interaction of grammar and other systems.
And this position is reiterated in Chomsky’s most recent paper [23] (p. 7):Derivations yield structures interpreted at two interfaces: conventionally termed SM and CI; postulation of interface levels is in fact superfluous; it is enough to say that extralinguistic systems access derivations.
Instead, from 2015 on, the SMT is formulated in terms of the simplest computational operation interacting with principles of minimal computation (computational efficiency).It is conventional to speak of two interfaces, the SM and ConceptualIntentional (CI) interfaces. While the device is convenient for exposition (and I’ll adopt it here), there is no need to postulate the interface levels; access can, in principle, take place at any stage of the computation.
The UCLA Lectures places this renewed emphasis on computational operations within the historical context of the Minimalist Program from its inception (p. 20):In the best case, phenomena would be explained by interaction of the simplest computational operation—Merge, with its two logically possible subcases. Internal Merge IM (automatically yielding “the copy theory of movement”) and External Merge EM—interacting with general principles of minimal computation MC. The Strong Minimalist Thesis SMT articulates this goal.[16] (p. 4)
The shift in focus from interface conditions (1992–2013) to the simplest computational operation is actually a return to the origins of modern generative grammar, where syntax is defined—for the first time in the study of language—as “the study of the principles and processes by which sentences are constructed in particular languages.”14Well, by the early 1990s it appeared to a number of people, myself included, that enough had been learned so that it might be possible to tackle the whole general problem of language on more principled grounds—that is, to actually seek genuine explanations, to formulate what was called the “Strong Minimalist Thesis”, which says that language is basically perfect, what’s called “approaching Universal Grammar from below”, starting by assuming (let’s see how far we can get) a perfect answer that would be a genuine explanation and how much can we explain this way. That’s what’s been called the Minimalist Program. Now the way to proceed with it is to start with the simplest computational operation and just let’s see how far you can go. Where do you run aground?
3. The Simplest Computational Operation
Passiveoptional: Structural analysis: NPAuxVNP Structural change: X_{1}X_{2}X_{3}X_{4} → X_{4}X_{2} + be + enX_{3}by + X_{1} 
Sentence  →  NP VP 
VP  →  Verb NP 
NP  →  T N 
Verb  →  Aux V 
Aux  →  C (M) (have + en) (be + ing) 
T  →  the, … 
N  →  thieves, police, … 
V  →  arrest, … 
In this formulation Merge specifies both the composition and labelling of syntactic objects, segmentation and classification (the classical discovery procedures of structuralist linguistics), which also results from the application of phrase structure rules. It differs from phrase structure rules in not specifying a linear order to the structures generated. Where α, β and {α, β} are syntactic objects, δ is clearly not, and {α, β} is merely the unlabeled counterpart to {δ, {α, β}}.18Given the numeration N, C_{HL} may select an item from N (reducing its index) or perform some permitted operation on the structures it has already formed. One such operation is necessary on conceptual grounds alone: an operation that forms larger units out of those already constructed; call it Merge. Applied to two objects α and β, Merge forms the new object γ. What is γ? γ must be constituted somehow from the two items α and β; the only alternatives are that γ is fixed for all α, β, or that it is randomly selected; neither is worth considering. The simplest object constructed from α and β is the set {α, β}, so we take γ to be at least this set, where α and β are the constituents of γ. Does that suffice? Output conditions dictate otherwise; thus, verbal and nominal elements are interpreted differently at LF and behave differently in the phonological component (see note 9). γ must therefore at least (and, we assume, at most) be of the form {δ, {α, β}}, where δ identifies the relevant properties of γ; call δ the label of γ.
As developments nine years later established, what seems to go wrong here is the assumption that to incorporate Move, the definition of Merge would have to become “more complex”.20One could define Merge differently, of course, allowing ‘internal Merger’. But there is no conceptual advantage in that; in fact, the definition is more complex.
So, when The Minimalist Program was published in 1995, it seemed clear that UG involved more than the simplest computational operation; and, moreover, that while Merge appeared to be a relatively simple operation, Move was a complicated one.So far, I have kept to the standard assumption that the operation Move selects α and raises it, targeting K, where α and K are categories constructed from one or more lexical items. But on general minimalist assumptions, that is an unnatural interpretation of the operation. The underlying intuitive idea is that the operation Move is driven by morphological considerations: the requirement that some feature F must be checked. The minimal operation, then, should raise just the feature F: we should restrict α in the operation Move α to lexical features. Let us investigate what happens if we replace the operation Move α by the more principled operation Move F, F a feature.(p. 262)
A Move F analysis of displacement/dislocation required another operation to account for the overt movement of α: “minimal piedpiping”. The formulations of Merge vs. Move maintain the dichotomy that existed between their precursors, phrase structure rules and transformations. Therefore, at that point, postulation of a strong minimalist thesis in terms of reducing UG to the simplest computational operation was demonstrably “too extreme to be seriously considered.”We are now tentatively assuming that if all features of some category α have been checked, then α is inaccessible to movement, whether it is a head or some projection. But if some feature F is, as yet, unchecked, α is free to move. Economy conditions exclude “extra” moves and anything more than the minimal piedpiping required for convergence. In covert movement, features raise alone. Procrastinate expresses the preference for the covert option.(p. 266)
How the phrase determined by F is calculated constitutes an extra step, designated as generalized piedpiping, which feeds Merge.First, what operations enter into this component of C_{HL}? One is indispensable in some form for any languagelike system: the operation Merge, which takes two syntactic objects (α, β) and forms K(α, β) from them. A second is an operation we can call Agree, which establishes a relation (agreement, Case checking) between an LI α and a feature F in some restricted search space (its domain). Unlike Merge, this operation is languagespecific, never built into specialpurpose symbolic systems and apparently without significant analogue elsewhere. We are therefore led to speculate that it relates to the design conditions for human language. A third operation is Move, combining Merge and Agree. The operation Move establishes agreement between α and F and merges P(F) to αP, where P(F) is a phrase determined by F (perhaps but not necessarily its maximal projection) and αP is a projection headed by α.(p. 101)
The third and final step happens with the realization that what had been considered to be two separate computational operations, Merge and Move, were in fact the same elementary operation applied in two distinct ways. As expressed in [10]:Plainly Move is more complex than its subcomponents Merge and Agree, or even the combination of the two, since it involves the extra step of determining P(F) (generalized “piedpiping”).(p.101)
As the simplest formulation for a recursive operation that yields the structured expressions of language, Merge constitutes the null hypothesis, which requires no explanation, in accord with the formulation of the SMT in this paper whereby there are no unexplained elements of UG. Therefore, it comes for free as well as being unconstrained (i.e., it applies freely).N[arrow] S[yntax] is based on the free operation Merge. SMT entails that Merge of a, b is unconstrained, therefore either external or internal. Under external Merge, a and b are separate objects; under internal Merge, one is part of the other, and Merge yields the property of “displacement,” which is ubiquitous in language and must be captured in some manner in any theory. It is hard to think of a simpler approach than allowing internal Merge (a grammatical transformation), an operation that is freely available. Accordingly, displacement is not an “imperfection” of language; its absence would be an imperfection.[footnote omitted] (p. 110)
4. Computational Efficiency
On page 6 “general considerations of computational efficiency” are characterized as “principles of efficient computation”, which are languageindependent (p. 9). Yet as [10] cautions, “the questions are empirical at every point, including the kinds of computational efficiency that FL selects” (p. 106).The PrinciplesandParameters approach opened the possibility for serious investigation of the third factor, and the attempt to account for properties of language in terms of general considerations of computational efficiency, eliminating some of the technology postulated as specific to language and providing more principled explanation of linguistic phenomena.
In effect, the first sentence merges the inclusiveness condition “operations forming complex expressions should consist of no more than a rearrangement of the objects to which they apply” and the NTC “not modifying them internally by deletion or (the) insertion of new elements.” As noted on page 13, the NTC “also entails the socalled copy theory of movement, which leaves unmodified the objects to which it applies, forming an extended object.” The underlying assumption is that the NTC holds for the elementary operation Merge, as [11] explains:One natural property of efficient computation, with a claim to extralinguistic generality, is that operations forming complex expressions should consist of no more than a rearrangement of the objects to which they apply, not modifying them internally by deletion or insertion of new elements. If tenable, that sharply reduces computational load: what has once been constructed can be ‘‘forgotten’’ in later computations, in that it will no longer be changed. That is one of the basic intuitions behind the notion of cyclic computation. The EST/Ymodel and other approaches violate this condition extensively, resorting to bar levels, traces, indices, and other devices, which both modify given objects and add new elements. A second question, then, is whether all of this technology is eliminable, and the empirical facts susceptible to principled explanation in accord with the ‘‘notampering’’ condition of efficient computation.
Thus, the copy theory would follow from either the Inclusiveness Condition or the NTC.NTC has always been assumed without comment for EM: there is, for example, no proposal that if V and NP are merged to form VP, then V is merged inside NP. Under SMT, it should hold for IM as well.(p. 10)
The question that arises is whether the NTC and a strict cycle condition are separate conditions of efficient computation. Depending on how the latter is formulated, both could bar the modification of syntactic objects created at earlier stages in a derivation; however, only the NTC prohibits modification via deletion.A Mergebased system will be compositional in general character: the interpretation of larger units at the interfaces will depend on the interpretation of their parts, a familiar observation in the study of every aspect of language. If the system is computationally efficient, once the interpretation of small units is determined it will not be modified by later operations—the general property of strict cyclicity that has repeatedly been found.
Strict Cycle Condition (SSC)
This formulation predates Move α and, therefore, assumes a formulation of (transformational) rules that is closer to that of the passive transformation cited above. Thus, in the case of this transformation, S would be a cyclic node.22 Basically, the idea is that operations cannot modify a subdomain of any syntactic object to which they are applying. In the first formulation of the Minimalist Program [2] strict cyclicity is recast as an extension condition such that “substitution operations always extend their target” (p. 23).23 Substitution operations here refer to the precursors of EM and IM, so in essence Merge always extends the syntactic objects to which it applies. This precludes inserting one syntactic object inside the other, in which case EM “always applies in the simplest possible form: at the root” [3] (p. 248), a conclusion that extends to overt IM (p. 254). This prohibition is also a consequence of the NTC, as noted in [11] where root is replaced by edge (p. 11):No rule can apply to a domain dominated by a cyclic node A in such a way as to affect solely a proper subdomain of A dominated by a node B which is also a cyclic node.
Thus, it would seem that strict cyclicity follows directly from the NTC and, therefore, is not by itself an independent principle of efficient computation.24Unless an element Z is an isolated element (an interjection, or frozen expression), hence of no interest here, its label W must have a feature indicating that Z can be merged. Under NTC, merge will always be to the edge of Z, so we can call this an edge feature EF of W.
And in [13], copy deletion is further clarified:So you have massive reduction in complexity if you just don’t pronounce extra copies. So computational efficiency compels the deletion of the copies.(p. 29)
See also [23], footnote 15. However, if the account of Merge in terms of item/context architecture is viable, then in this case it is the formulation of the simplest computational operation (Merge) rather than third factor principles of computational efficiency that accounts for the phenomenon of displacement.In externalization, all but one of a copy set is deleted by a third factor principle of computational efficiency.(p. 11)
5. Beyond the Simplest Computational Operation?
 [1]
 (i) It must be rich enough to overcome the problem of poverty of stimulus (POS), the fact that what is acquired demonstrably lies far beyond the evidence available(ii) It must be simple enough to have evolved under the conditions of human evolution(iii) It must be the same for all possible languages, given commonality of UG.
The question then is what operations are required in addition to Merge and how do they qualify as “simplest computational operations”.The core system of language might be largely immune to the learnability condition [1i], which has played a prominent role in the generative enterprise. The variability problem [1iii] will be assigned to an ancillary system, the amalgam of language and some SM system. [footnote omitted] The conundrum will then reduce to satisfying the evolvability condition [1ii]. That problem will be overcome, at least for the core computational system of language–our topic here–to the extent that the structures of Ilanguage are generated by the simplest operations. The Strong Minimalist Thesis SMT sets this outcome as a prime goal of the theory of language.
Where the string of adjective phrases following someone is potentially unbounded (hence the …), with no structural relation between them. But each adjective phrase modifies the same noun someone. In terms of an item/context architecture, this could be described as multiple items (the adjective phrases) that have the same context (the noun someone, which each phrase modifies). Thus, coordination would be a case of a single context that has multiple items, in contrast to displacement where a single item is assigned multiple contexts. But given that these adjective phrases have no structural relation to one another, they could not be constructed by Merge on a single plane (i.e., in the same dimension). Rather, Merge would have to be able to form multidimensional structures where the noun is accessible to each adjective phrase and what appears to be a sequence of adjective phrases is the result of linearization for the SM interface.34 Whether this analysis can be extended to the wider range of more complicated cases remains to be determined.I met someone young, happy, eager to go to college, tired of wasting his time, …
Funding
Conflicts of Interest
Notes
1  See [4] for an overview of the minimalist program as presented in the book. As Chomsky makes clear in the 1992 paper, the minimalist program is formulated in terms of the Principles and Parameters framework, thereby incorporating the perspective and some of the proposals of earlier work in generative grammar. The introduction to the 20th anniversary edition of the 1995 collection describes it as “a seamless continuation of pursuits that trace back to the origins of generative grammar, even before the general biolinguistics program, as it is now often called, began to take shape in the 1950s” (p. vii). See also [5] for a discussion of some fundamental ideas the minimalist program is investigating that can be found in some form in the linguistic theory that precedes it. 
2  Page references are to the version that occurs in [3]. 
3  See [8] for some discussion. 
4  These include “acquisition, processing, neurology, language change, and so on” (p. 96). 
5  Language design has been a constant focus in discussion of the minimalist program from the outset. It predates the proposal of a minimalist program [2] in Chomsky’s “Economy of Derivation and Representation” [12]. It is worth noting that the topic as currently discussed first appears in Chomsky & Lasnik’s “Filters and Control” 1977 [12] (p. 434) when they state:

6  The words perfect or perfection do not occur in the initial formulation of a minimalist program [2], but the second paper on the minimalist program two years later [6] begins with a speculation about whether “language is something like a “perfect system.” This is repeated in the introduction to [3] the following year, which begins:
Chomsky notes that in contrast to question A, question B “has an exact answer, though only parts of it can be surmised in the light of current understanding about language and related cognitive systems”, an appraisal which accurately assesses the current situation. Nonetheless, it is in this context that the speculation arises:
So pretty much from the outset, the Minimalist Program is concerned with exploring the possibility that the system of language not only functions in an optimal way, but may be in fact a perfect system in the way it interfaces with other cognitive systems. 
7  Or “genuine explanations” in the terminology of Chomsky 2021a [13]. 
8  GPs abbreviates “generative procedures”. 
9  The designation UP for “universal phonetics” later becomes PF for “phonetic form”. 
10  Cf. The Functional Relatedness constraint in [19]. 
11  The term “bare output condition” designates conditions external to UG, in contrast to UG internal constraints such as filters and conditions on derivations. Ref. [3] goes on to say: “The problems are nevertheless empirical, and we can hope to resolve them by learning more about the language faculty and the systems with which it interacts. We proceed in the only possible way: by making tentative assumptions about the external systems and proceeding from there” (pp. 222–223). 
12  As a striking demonstration, consider the over five million variants of the first sentence of Jane Austen’s Pride and Prejudice enumerated in [20] that might have appeared in its place had Austen not been such a brilliant writer. 
13  The last reference to interface conditions in statements of the SMT occurs in the introduction to the 20th anniversary edition of The Minimalist Program (quoted above), but is not mentioned in the formulation given in “Problems of Projection: Extensions” published the same year (quoted below). 
14  
15  For a detailed review of the successive changes the passive transformation underwent from its initial formulation to its penultimate simplification as the operation Move α (see [26]). In particular, the analysis in which the subject of a corresponding active sentence is displaced (or demoted) in the passive no longer applies. But see Collins 2005 for an attempt to reconstruct the full Syntactic Structures analysis within current analyses of clause structure [27]. 
16  Transformations whose output can be filtered in this way by the phrase structure rules are structurepreserving, as discussed in Emonds’s 1970 Ph.D. dissertation [28] and his 1976 book [29], the Structure Preserving Hypothesis (SPH), as it came be called. In the fourth chapter of [3] the SPH is criticized for introducing “an unwanted redundancy” on the grounds that the movement of a constituent creates a position without there having to be an empty position into which the constituent moves. However, Emonds’s formulations of the SPH do not depend on this substitution analysis. 
17  A numeration consists of a set of lexical items each associated with a numerical index that indicates how many times a particular item has been selected from the lexicon. Thus, in the passive sentence discussed above, the article the would have the index 2 while all the other lexical items in the array would have an index of 1. Each time an item in the numeration is affected by an operation of C_{HL}, its index is reduced by one. At the end of a derivation, each item in the numeration should have an index of zero. The device of a numeration disappears from discussion after chapter 4 of The Minimalist Program. The perhaps obvious alternative is a lexical array where a single lexical item can occur more than once. The goal of a derivation would then be a lexical array that is empty. If, alternatively, the lexical array constitutes the initial workspace for a derivation, then the goal would be to reduce the workspace to a single syntactic object, the linguistic expression derived from the interaction between C_{HL} and the lexicon. See below for further discussion. 
18  Starting with [1], it is assumed that labels of syntactic objects are determined by general algorithms, and therefore that Merge creates a set of syntactic objects. Ref. [1] also introduces the suggestion that the output of Merge is either a set or a pair (hence SetMerge or PairMerge). Whether this undermines the goal of reducing UG to the simplest computational operation remains to be determined. For the moment we will consider Merge to be an operation that produces sets. 
19  Note that phrase marker here can’t refer to a set of strings as originally defined given the elimination of phrase structure rules that produce such sets. Rather, it can only refer to the hierarchically structured syntactic object that is produced by Merge. 
20  It is important to note that these references to merger in the description of Move are eliminated in chapter 4 of The Minimalist Program, which incorporates [6], much of it verbatim. 
21  
22  For a detailed discussion of this formulation and how its empirical effects can be derived from independently motivated general principles, see [19]. 
23  For discussion of some empirical effects of the extension condition, see [34], which concludes (p. 96):
It should be noted that this was written when Merge and Move were treated as independent operations. 
24  See [35] for further discussion. 
25  In recent work (e.g., [22]) this formulation is designated as MERGE, to distinguish it from earlier formulations. In earlier work, a workspace containing syntactic objects constructed by Merge is tacitly assumed, if not mentioned—the exception being [36] where Merge is presented as an operation that maps a workspace onto a modified workspace. The earliest reference to “workspace” in Chomsky’s writings occurs in [1] where the term is equated with “active memory” (p. 106). 
26  If Merge is formulated as a mapping between workspaces, then presumably Merge does not operate on the Lexicon directly. Placing the Lexicon in the workspace creates a situation where the entire content of the Lexicon would be accessible to Merge at any point in the derivation, undermining efficient computation (which presumably is limited to the lexical items of the linguistic expression under analysis). 
27  See [22] for discussion of how for natural language this could lead to any kind of island violation. Ref. [22] proposes a “resource restriction” limiting new syntactic objects in a derived workspace to one (cf. [36] where it is first mentioned as an observation), thereby forcing the elimination of the unmerged copies of the newly merged elements. See [22] for discussion of how this restriction prohibits, in principle, operations such as Parallel Merge, Sidewards Merge, and Late Merge. However this would also follow from a formulation of Merge where elements merged do not remain in the workspace as unconnected terms. 
28  IM also specifies the composition of linguistic expressions, so composition and displacement are not disjoint properties. 
29  This terminology originates with JeanRoger Vergnaud (p.c. January 2004), see [37]. 
30  Refs. [13,31] propose a separate operation Form Copy that applies not only to the output of IM for displacement, but also to a range of other syntactic phenomena, including obligatory control, acrosstheboard deletion, and parasitic gaps. The analysis is grounded in a new idea that the SMT has an “enabling function” as well as a restrictive function.

31  How this actually applies to the obvious counterexamples, virtually every case of ellipsis, where deletion of phonetic material is entirely optional, remains to be determined. 
32  Given the Access Restriction proposed here, the empirical effects of the Resource Restriction follow as a consequence. However, the Resource Restriction does not subsume the NTC. 
33  This leaves open the large question of locality constraints on IM and thus the anlaysis of the phenomenon of syntactic islands. The Phase Impenetrability Condition (see [8,14]) seems at this point to be a UGspecific constraint on derivations, though in limiting the scope of IM it might be thought of as a principle of minimal computation. 
34  Refs. [13,22,23] propose this ndimensional analysis for adjuncts (including unbounded unstructured coordination (of which the term itself is an example)), employing another operation PairMerge to form these structures. Why these structures could not be constructed with SetMerge is not entirely clear. Refs. [13,23] propose another operation FormSequence for generating coordinate structures, described as “a limited departure from SMT”. Whether this operation is in addition to or a replacement of PairMerge is unclear. 
References
 Chomsky, N. Minimalist inquiries: The framework. In Step by Step: Essays in Minimalist Syntax in Honor of Howard Lasnik; Martin, R., Michaels, D., Uriagereka, J., Eds.; M.I.T. Press: Cambridge, MA, USA, 2000; pp. 89–155. [Google Scholar]
 Chomsky, N. A minimalist program for linguistic theory. In The View from Building 20: Essays in Linguistics in Honor of Sylvain Bromberger; Hale, K., Keyser, S.J., Eds.; M.I.T. Press: Cambridge, MA, USA, 1993; pp. 1–52. [Google Scholar]
 Chomsky, N. The Minimalist Program; M.I.T. Press: Cambridge, MA, USA, 1995. [Google Scholar]
 Freidin, R. Review article on The Minimalist Program by Noam Chomsky. Language 1997, 73, 571–582. [Google Scholar] [CrossRef]
 Robert, F.; Lasnik, H. Some roots of minimalism in generative grammar. In The Oxford Handbook of Linguistic Minimalism; Boeckx, C., Ed.; Oxford University Press: New York, NY, USA, 2011; pp. 1–26. [Google Scholar]
 Chomsky, N. Bare phrase structure. In Evolution and Revolution in Linguistic Theory: Essays in Honor of Carlos P. otero; Campos, H., Kempchinsky, P., Eds.; Georgetown University Press: Washington, DC, USA, 1995; pp. 51–109. [Google Scholar]
 Chomsky, N. The Minimalist Program, 20th ed.; M.I.T. Press: Cambridge, MA, USA, 2015. [Google Scholar]
 Chomsky, N. Derivation by phase. In Ken Hale: A Life in Language; Kenstowicz, M., Ed.; M.I.T. Press: Cambridge, MA, USA, 2001; pp. 1–52. [Google Scholar]
 Chomsky, N. What Kind of Creatures Are We? Columbia University Press: New York, NY, USA, 2016. [Google Scholar]
 Chomsky, N. Beyond explanatory adequacy. In Structures and Beyond; Belletti, A., Ed.; Oxford University Press: Oxford, UK, 2004; pp. 104–131. [Google Scholar]
 Chomsky, N. Approaching UG from below. In Interfaces + Recursion = Language?: Chomsky’s Minimalism and the View from SyntaxSemantics; Sauerland, U., Gärtner, H.M., Eds.; Mouton de Gruyter: Berlin, Germany, 2007; pp. 1–30. [Google Scholar]
 Chomsky, N.; Lasnik, H. Filters and control. Linguist. Inq. 1977, 11, 425–504. [Google Scholar]
 Chomsky, N. Genuine explanation. In Issues in Comparative MorphoSyntax and Language Acquisition; Bocci, G., Botteri, D., Manetti, C., Moscati, V., Eds.; forthcoming.
 Chomsky, N. On phases. In Foundational Issues in Linguistic Theory: Essays in Honor of JeanRoger Vergnaud; Freidin, R., Otero, C.P., Zubizarreta, M.L., Eds.; M.I.T. Press: Cambridge, MA, USA, 2008; pp. 133–166. [Google Scholar]
 Chomsky, N. Problems of projection. Lingua 2013, 130, 33–49. [Google Scholar] [CrossRef]
 Chomsky, N. Problems of Projection: Extensions. In Structures, Strategies and Beyond: Studies in Honour of Adriana Belletti; Di Domenico, E., Hamann, C., Matteini, S., Eds.; John Benjamins: Amsterdam, The Netherlands, 2015; pp. 1–16. [Google Scholar]
 Chomsky, N.; Lasnik, H. The theory of principles and parameters. In Syntax: An International Handbook of Contemporary Research; Jacobs, J., van Stechow, A., Sternefeld, W., Vennemann, T., Eds.; Walter de Gruyter: Berlin, Germany, 1993; pp. 506–569. [Google Scholar]
 Chomsky, N. Knowledge of Language: Its Nature, Origin, and Use; Praeger: New York, NY, USA, 1986. [Google Scholar]
 Freidin, R. Cyclicity and the theory of grammar. Linguist. Inq. 1978, 9, 519–549. [Google Scholar]
 Freidin, R. Adventures in English Syntax; Cambridge University Press: Cambridge, UK, 2020. [Google Scholar]
 Chomsky, N. Cartesian Linguistics: A Chapter in the History of Rationalist Thought, 3rd ed.; Cambridge University Press: Cambridge, UK, 2009. [Google Scholar]
 Chomsky, N. The UCLA Lectures. LingBuzz. 2020. Available online: https://ling.auf.net/lingbuzz/005485 (accessed on 21 November 2021).
 Chomsky, N. Minimalism: Where Are We Now, and Where Can We Hope to Go. To Appear in Genko Kenkyu. 2021. Available online: https://www.youtube.com/watch?v=X4F9NSVVVuw (accessed on 21 November 2021).
 Chomsky, N. Syntactic Structures; Mouton: The Hague, The Netherlands, 1957. [Google Scholar]
 Freidin, R. Syntactic Structures: A Radical Appreciation. LingBuzz. 2020. Available online: https://ling.auf.net/lingbuzz/004996 (accessed on 21 November 2021).
 Freidin, R. Conceptual shifts in the science of grammar 1951–1992. In Noam Chomsky: Critical Assessments; Otero, C.P., Ed.; Routledge: London, UK, 1994; Volume 1, pp. 653–690. [Google Scholar]
 Collins, C. A smuggling approach to the passive in English. Syntax 2005, 8, 81–120. [Google Scholar] [CrossRef]
 Emonds, J. Root and Structure Preserving Transformations. Ph.D. Thesis, Massachusetts Institute of Technology, Cambridge, MA, USA, 1970. [Google Scholar]
 Emonds, J. A Transformational Approach to English Syntax: Root, StructurePreserving, and Local Transformations; Academic Press: New York, NY, USA, 1976. [Google Scholar]
 Freidin, R. Chomsky’s linguistics: The goals of the generative enterprise. Language 2016, 92, 671–723. [Google Scholar] [CrossRef]
 Chomsky, N. Some notes on economy of derivation and representation. In Principles and Parameters in Comparative Grammar; Freidin, R., Ed.; M.I.T. Press: Cambridge, MA, USA, 1991; pp. 417–454. [Google Scholar]
 Chomsky, N. Three factors in language design. Linguist. Inq. 2005, 36, 1–22. [Google Scholar] [CrossRef]
 Chomsky, N. Conditions on transformations. In A Festschrift for Morris Halle; Anderson, S., Kiparsky, P., Eds.; Holt, Rinehart and Winston: New York, NY, USA, 1973; pp. 232–286. [Google Scholar]
 Freidin, R. Cyclicity and minimalism. In Working Minimalism; Epstein, S.D., Hornstein, N., Eds.; M.I.T. Press: Cambridge, MA, USA, 1999; pp. 95–126. [Google Scholar]
 Freidin, R. Cyclicity in syntax. In Oxford Research Encyclopedia of Linguistics; Aronoff, M., Ed.; Oxford University Press: Oxford, UK, 2017; Available online: https://oxfordre.com/linguistics/view/10.1093/acrefore/9780199384655.001.0001/acrefore9780199384655e319?rskey=MlkDK4&result=1 (accessed on 21 November 2021).
 Bobaljik, J. In terms of Merge: Copy and Head movement. In Papers in Minimalist Syntax; Pensalfini, R., Ura, H., Eds.; MITWPL: Cambridge, MA, USA, 1995; Volume 27, pp. 41–64. [Google Scholar]
 Vergnaud, J.R. Primitive Elements of Grammatical Theory: Papers by Jean Roger Vergnaud and His Collaborators; McKinneyBock, K., Zubizarreta, M.L., Eds.; Routledge: New York, NY, USA, 2014. [Google Scholar]
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. 
© 2021 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Freidin, R. The Strong Minimalist Thesis. Philosophies 2021, 6, 97. https://doi.org/10.3390/philosophies6040097
Freidin R. The Strong Minimalist Thesis. Philosophies. 2021; 6(4):97. https://doi.org/10.3390/philosophies6040097
Chicago/Turabian StyleFreidin, Robert. 2021. "The Strong Minimalist Thesis" Philosophies 6, no. 4: 97. https://doi.org/10.3390/philosophies6040097