Daniele GODARD (Daniele.Godard@linguist.jussieu.fr)
Fri, 24 Jan 1997 19:24:16 +0100 (MET)
In the recent discussion about adverbs as complements, two questions have been raised concerning the analysis of adverbs as complements in French (the 'ackems' group; I note that the initial of my name gets assimilated!): 1. The unicity argument: We argue that the fact that only one negative adverb can occur in the finite VP is a property which negative adverbs share with complements, and an argument in favor of their being analyzed as complements. A. The facts. Some doubt has been cast about the data. Let us go back to them: (1) * Paul ne lit pas pas Proust (2) * Paul n'a pas pas lu Proust (3) * Paul n'a nullement pas lu Proust These sentences are bad; the only situation where one can pronounce them is if one quotes or echoes a previous sentence: speaker A: Paul n'a pas lu Proust speaker B: Paul n'a pas "pas lu Proust". But then, everything is possible in these conditions. The fact that 'nullement' is phonetically distinct from 'pas' may make (3) appear better , but doesn't change the fact that it is bad. I have checked with J-C. Milner who uses 'nullement' in a natural way (this is not the case for everybody; 'nullement' is somewhat archaic, or refined). Note that double negations are possible, including one with 'pas' and a negative nominal, with a double negation interpretation (see work by F. Corblin): (4) Paul n'a pas lu aucun livre The status of (4) is different from that of (1)-(3). B. Why is this an argument? As stressed by Bob in his summary, it has been noted that complements differ from adjuncts in that they are constrained to be 'unique' (only one of its kind). This is what we had in mind, and I'm not sure it is what Tibor had in mind, when he discussed the argument (I think he had something like cancellation in mind?). So, if the negative adverb is included in the complements, the fact that it appears only once will follow from the general constraint on the unicity of complements. It is true that the constraint has to be made more precise; nevertheless, it seems to be a sound observation. 2. Why we now restrict the adverb insertion LR to negative adverbs? Bob is exactly right when he says that we see this as an empirical problem: we restrict the rule to those adverbs for which we think we have arguments in favor of their complementhood. 3. Why A and G are discrete about scope and extraction: simply because the question is not solved, and we wanted to have as simple an argument as possible. As regards the relative scope of two adverbs, A and B's position is the following; remember that adverbs are either adjoined to a head (the head-adjunct schema) or at the same level as complements (a head-compl-adj schema): - the adverb should have (at least) scope over the domain it c-commands; - if the two adverbs are at the same level, then the scope if from left to right, that it, the adv to th eleft has scope over the adv to the right (Horn): (5) Jean a deliberement frequemment repondu aux questions J. has deliberately frequently answered the questions (6) Jean a deliberement repondu frequemment aux questions (7) ? Jean a repondu deliberement frequemment aux questions (8) Jean a frequemment deliberement repondu aux questions (9) Jean a frequemment repondu deliberement aux questions (10) Jean a repondu frequemment deliberement aux questions These sentences are not ambiguous: (5)-(7) convey that it is on purpose (deliberement) that J. frequently answered the questions, and (8(10) that it was frequently that J. decided to answer the questions. Now, if we have (11), the sentence is ambiguous: (11) Jean repond frequemment aux questions deliberement In our analysis, this follows from the two configurations to whihc this sentence can be assigned: - either 'deliberement' is at the same level as the complement and the adv frequemment; in this case, frequemment has scope over it (it is to the left); - or, deliberement is adjoined to the VP; if so, it is higher than the VP including frequemment; hence, it has scope over it. I don't see how these data and analysis can be made as argument in favor or disfavor of the status of adverbs as complements. Daniele Godard
This archive was generated by hypermail 2.0b3 on Fri Dec 18 1998 - 20:36:03 PST