| Description |
A long tradition takes word order differences within one and the same language and across languages to be a consequence of movement.
This tradition has treated together movements which have interpretive effects at both PF (word order change) and LF (meaning change), and movements which have effects at PF (word order change) but no clear interpretive effects at LF. Among the latter, head-(or phrasal)-movements within one and the same language (say, Italian) or across languages (French-English), and especially the phrasal movements that have been taken to derive the different head-initial and head-final word orders of languages (Kayne 1994, Biberauer,Holmberg and Roberts 2014, Cinque 2017).
More recently Chomsky (1995,2000,2013) has made the not implausible suggestion that, given the architecture of UG, internal Merge (movement) in narrow syntax must have interpretive effects at both interfaces (LF and PF). Any putative movement that has an effect on the PF side but no effect on the LF side should be expunged from narrow syntax. Consequently, any word order difference with no LF effect should be confined to the PF interface.
The same would apply to putative movements that have an effect on the LF side but no apparent effect on the PF side, like the traditional LF movements of quantifiers, in cases like I will force you to marry no one. They too should be expunged from narrow syntax, unless they can be shown to have a component that affects the PF side. Excluding pronunciation of the low copy of movement, one possibility is to take ‘LF movement’ to actually be overt movement of the quantifier in narrow syntax (Kayne 1998), followed by remnant VP-preposing (the latter an interesting residue of meaningless movement, apparently related to the head-initial/head-final character of the language (cf. English vs. German). I return to this residue and to possible further evidence that some meaningless movement should possibly be countenanced.
Chomsky’s position in 1995 was to take Kayne’s (1994) LCA “to be a principle of the phonological component” (p.340), but this position was later abandoned as it implies some meaningless movement “to yield the proper hierarchies” (Chomsky 2004:110 and note 27) that determine the different linear orders of languages.
The current lack of an explicit mechanism of externalization risks taking us back to Chomsky’s (1965) dismissal of set-systems, as “[n]o proponents of a set-system has given any indication of how the abstract underlying unordered structures are converted into actual strings with surface structures.” (p.125).
Though, of course, it is perfectly possible that some externalization mechanisms will be found, the task is anything but trivial. To assume “that the physics of speech demand that linguistic units must be pronounced sequentially in time, giving rise, in this case, to just two options: the head either precedes or follows its complements” (Eguren, Fernández-Soriano and Mendikoetxea (2016:12, who cite many sources taking this position) seems far from being sufficient. It falls short of accounting for a number of generalizations. I will in particular discuss two interrelated generalizations concerning the DP. One concerns the attested vs. unattested orders of modifiers (Cinque 2005) and the other the possible vs. impossible interpretations of DP-internal ellipses (the generalization being that “ellipsis can target a nominal modifier only if all constituents below it are also elided” Cinque 2012). Under the assumption that ellipsis (non-pronunciation) of a constituent depends on the prior movement of that constituent to a left-peripheral position (Jayaseelan 1990, Rizzi 1994, Johnson 2001, Ntelitheos 2004, Kayne 2006, 2012; among others), the two generalizations can be argued to follow from the same mechanism, the way the NP moves, possibly pied piping constituents containing it. If so, the LCA can be retained, whether in narrow syntax (in a possibly modified form - Kayne 2018), or PF.
It is not clear how recent direct linearization theories for the DP (Kremers 2009, where some movements are still assumed, and Hall 2015,Chapter 5) could express the generalization uniting word order and ellipsis.
|
