A brief history of the passive transformation
Consider for example, the analysis of passive sentences and the passive transformation, perhaps the most central transformation within early transformational analysis. In SS we find the following formulation:
The passive transformation as formulated in (1) performs several distinct operations. It permutes an NP subject in front of an auxiliary phrase20 with an NP following the verb (usually construed to be the object of the verb). In addition it adjoins the passive auxiliary be+en to the right of the auxiliary phrase, and also adjoins the grammatical formative by to the left of the permuted subject NP. The rule is stipulated to be optional, indicating that it need not apply even if its structural description is met in the course of a derivation. Given this formulation, an active sentence and its corresponding passive counterpart would have the same underlying deep structure. Thus the transformational analysis expressed the relation between active and passive sentential structures in terms of a common deep structure.
In Aspects Chomsky tries to improve on the analysis of the passive transformation by proposing that the passive wh-phrase constituted an instance of a manner adverbial. The motivation comes from an observation that passivization seemed possible only for verbs that could occur with manner adverbs.
These observations suggest that the Manner Adverbial should have as one of its realizations a ‘dummy element’ signifying that the passive transformation must obligatorily apply. That is, we may have the rule (2) as a rewriting rule of the base and may formulate the passive transformation so as to apply to strings of the form (3), with an elementary transformation that substitutes the first NP for the dummy element passive and places the second NP in the position of the first NP:
(2)
(3)
Thus the wh-phrase would exist in underlying structure, thereby eliminating the need to insert a language-specific grammatical element (by) using the passive transformation. The analysis also solved the problem of the derived constituent structure of the by-phrase, which in SS had to be specified by an ad hoc rule of derived constituent structure. It assumed that markers such as passive were ‘drawn from a fixed, universal, language independent set’ (p. 223).
The analysis of the passive auxiliary be+en is not discussed explicitly in Aspects. Note that the description of the passive transformation quoted above does not mention the insertion of the auxiliary as part of the transformation. This leaves open the possibility that the passive auxiliary is introduced into the derivation via a phrase structure rule. Yet it is not clear from the Aspects phrase structure analysis of Aux that the passive auxiliary could be introduced via a phrase structure rule. The rewrite rule for Aux given in (57.xvi) on page 107 is essentially the same as the one given in SS on page 111 (i.e. Aux→C(M) (have+en) (be+en)) except that C is replaced by T (for tense) and the two aspectual auxiliaries are covered by a single designation Aspect’. However it is unlikely that the passive auxiliary was assumed to be introduced via a phrase structure rule given the discussion in SS concerning the problems created by such an analysis.
In ‘Remarks on nominalization’ (written in 1967, published in 1970; henceforth ‘Remarks’), the analysis of passives and the passive transformation undergoes an even more radical revision. In ‘Remarks’ Chomsky considers the correspondence between sentences and nominals, as illustrated in (4)–(5).
(4)
(5)
Earlier work on nominalizations (i.e., Lees 1960) assumed that the nominals (5a,b) were derived via a nominalization transformation which applied to the sentential forms (4a, b) respectively. Chomsky rejects the transformational analysis of nominals on the grounds that such transformations do not have the general character of other standard transformational rules. They are not productive—that is, not all verbs have a corresponding derived nominal. Furthermore, the relation between a derived nominal and its associated verb is idiosyncratic.
Consider, for example, such nominals as laughter, marriage, construction, actions, activities, revolution, belief, doubt, conversion, permutation, trial, residence, qualifications, specifications and so on, with their individual ranges of meaning and varied semantic relations to the base forms. There are a few subregularities that have frequently been noted, but the range of variation and its rather accidental character are typical of lexical structure. To accommodate these facts within the transformational approach (assuming, as above, that it is the grammatical relations in the deep structure that determine meaning) it is necessary to resort to the artifice of assigning a range of meanings to the base form, stipulating that with certain semantic features the form must nominalize and with others it cannot. Furthermore, the appeal to this highly unsatisfactory device, which reduces the hypothesis that transformations do not have semantic content to near vacuity, would have to be quite extensive. (p. 19)
In addition, nominalization transformations, unlike other transformations, require the ability to change categorial analysis—for example, changing a verb into a noun and correspondingly a sentential structure into an NP. However, given the theory of syntactic features developed in Aspects, Chomsky is able to formulate the lexicalist hypothesis which can accommodate the analysis of derived nominals in the lexicon and thereby achieve a simplification of the transformational component. The lexicalist hypothesis prohibits feature changing by transformation as well as transformations whose application depends on idiosyncratic features of lexical items. It assumes that idiosyncratic properties of individual lexical items are captured in the lexicon.
Though the nominal (5b) is not transformationally derived from the sentence (4b), it s transformationally related to the nominal (5a) in much the same way that the passive construction in (4b) is transformationally related to the corresponding active construction (2a). The semantic relation between city and destruction, as well as enemy and destruction, is identical in both nominals even though their surface syntax is different. Chomsky analyses the nominal (4b) as derived from an underlying structure resembling (4a).
(6)
A indicates a base-generated empty category (NP). Thus in (6) the city bears the relation object of the nominal and the enemy bears the relation subject of the nominal in both (5a) and (5b). These grammatical relations are identical to those in the corresponding sentential forms (4a,b). Putting aside the insertions of ‘s in front of the nominal and of after the nominal, (5b) is derived from (6) via transformational operations that substitute the NP the enemy for ∆ and fill the empty subject position with the city (via another substitution operation). Chomsky establishes the independence of the two operations by pointing out that Agent postposing may apply without the application of NP preposing (even when it could apply as in (5c)), and that more generally Agent postposing does not require the presence of an object as illustrated by the derivation of (7a) from (7b).
(7)
If the wh-phrase in (7b) were missing, then Agent postposing could not apply and John’s offer would be derived. Furthermore, if the absence of the wh-phrase in derived nominals is treated as the absence of the phrase in underlying structure, given its optionality in phrase structure, then NP preposing could apply in the absence of Agent postposing, as illustrated in (8a) (derived from (8b)).
(8)
If NP preposing does not apply to (8b), then the appointment of John to the committee is derived.
Chomsky characterizes the two rules that apply to nominal constructions as follows: ‘Agent-postposing is simply a generalization of one of the components of the passive transformation. NP-preposing is similar to, and may fall under a generalization of, the other component’ (1970:43). Note that he still assumes the existence of a passive transformation even though he has just demonstrated that the two movement components of the rule that applies to sentential structures apply independently in nominal constructions. Presumably this is because he still assumes that the passive auxiliary must be inserted via the transformation given his arguments in SS against base-generating the passive auxiliary. Thus the analysis in ‘Remarks’ identifies two subparts of the passive transformation as grammatical transformations with respect to nominals and as ‘elementary’ transformations (in the sense of LSLT) with respect to sentential structures. A similar analysis occurs in Hasegawa 1968 which identifies the postposing of agents (i.e., Agent-postposing) in gerundive nominals (e.g., the polishing of the lenses by the workers) as an independent grammatical transformation.
With hindsight, we now know that the analysis in ‘Remarks’ constitutes the first important step towards a more general theory of transformations that does not allow the formulation of construction-specific transformations like Tpassive. It establishes that the movement component of the passive transformation is essentially a pair of substitution operations in which a lexical NP is substituted for an empty NP. The next step occurs with the realization that the rule of NP-preposing is actually an instance of a more general rule ‘Move NP’—or more accurately, Substitute NP. Thus much of the work of the early 1970s was directed towards an analysis of NP movement which generalized the NP preposing component of the passive transformation to instances of subject-to-subject raising, as illustrated in (9), where ti indicates the underlying position of the co-indexed matrix subject and [α indicates a clause boundary
(9)
(Note that the matrix predicate in each case is distinct—a predicative adjective in (9a), a verb in (9b) and a passive predicate in (9c).) Given a rule Move NP, the structural description for the transformation can be reduced to the bare essentials.
(10)
Only the movement site and the landing site are given as constant terms. The syntactic distance between the two positions must be mediated by other conditions which are not stated as part of the transformation—for example, the framework of conditions developed in Chomsky 1973, 1976. As it turned out, the development of a system of conditions on rules of grammar made it possible to formulate the more general theory of transformations, under the assumption that these conditions constituted general principles (i.e., of UG).
