Answer ellipsis (= answer fragments) is a type of ellipsis that occurs in answers to questions. Answer ellipsis appears very frequently in any dialogue, and it is present in probably all languages. Of the types of ellipsis mechanisms, answer fragments behave most like sluicing, a point that shall be illustrated below.
Standard instances of answer ellipsis occur in answers to questions. A question is posed, and the answer is formulated in such a manner to be maximally efficient. Just the constituent that is focused by the question word is uttered. The elided material in the examples in this article is indicated using a smaller font and subscripts:
This sort of data could easily be expanded. An answer fragment is possible for any constituent that can be questioned using a question word. An important aspect of the elided material of answer ellipsis is that it usually does not correspond to a constituent. This fact is problematic for theories of ellipsis, a point which is examined below.
Answer ellipsis behaves curiously in a couple of noteworthy ways. The answer fragment should not, for instance, encompass more than the focused constituent:
Either just the focused constituent (i.e. the constituent that is focused by the question word in the question) or the entire sentence must appear as the answer. If an intermediate constituent appears, the answer is unacceptable.
Another noteworthy aspect of answer ellipsis is that a negation can be part of the elided material. Answer ellipsis is like sluicing in this regard, but unlike gapping, stripping, VP-ellipsis, and pseudogapping, e.g.
These data demonstrate that answer ellipsis and sluicing have something important in common that distinguishes them from other ellipsis mechanisms.
Theoretical accounts of answer ellipsis are faced with the same basic problem that challenges the accounts of other ellipsis mechanisms. This problem revolves around the fact that the elided material often does not form a constituent in surface syntax. The following trees illustrate the problem. The tree on the left is a constituency-based tree of a phrase structure grammar, and the tree on the right is the corresponding dependency-based tree of a dependency grammar:
In both trees, the elided material I said, which is indicated using the lighter font shade, does not form as a constituent. In other words, it does not qualify as a complete subtree. The theory of ellipsis is therefore challenged to produce an analysis of such cases that can account for the fact that ellipsis appears to be eliding non-constituent units.
One prominent means of dealing with this problem is to assume that the answer fragment is first moved out of an encompassing constituent so that this encompassing constituent can then be elided. The following tree illustrates such an analysis in a phrase structure grammar:
The object nothing is moved to the left out of the constituent S in such a manner that S (the lower S) can then be elided. This sort of analysis allows one to preserve the assumption that ellipsis mechanisms (in this case answer ellipsis) are eliding constituents. A constituent-based theory of syntax can therefore be maintained.
Analyses of fragment answers that incorporate movement of the remnant of ellipsis in the manner just sketched have a way of accounting for two important facts documented in the literature. First, in languages that require preposition pied-piping such as German, fragment answers that retain the preposition (pied-pipe it, on the movement analysis) are judged significantly more acceptable than those that do not retain it, a phenomenon related to what is known as the Preposition-Stranding Generalization (or Merchant's Generalization). Second, fragment answers appear to be sensitive to locality restrictions on movement, including islands:
An alternative analysis takes the catena as the fundamental unit of syntactic analysis and assumes that answer ellipsis is eliding catenae. The catena is closely associated with dependency-based theories of syntax. It is defined as any word or any combination of words that is continuous with respect to dominance. Given this definition, any subtree of a tree qualifies as a catena. In the example above, the word combination I said qualifies as a catena in surface structure (without movement). Movement is hence not required to account for the fact that ellipsis appears to be eliding non-constituent units. The elided units are catenae, and as such they are clearly defined units of syntactic analysis.
Two additional complex examples further illustrate how a catena-based analysis of answer fragments works:
While the elided material shown in light gray certainly cannot be construed as a constituent, it does qualify as a catena (because it forms a subtree). The following example shows that even when the answer contains two fragments, the elided material still qualifies as a catena:
Such answers that contain two (or even more) fragments are rare in English (although they are more common in other languages) and may be less than fully acceptable. A movement analysis of this answer fragment would have to assume that both Susan and Larry have moved out of the encompassing constituent so that that constituent can be elided. The catena-based analysis, in contrast, does not need to appeal to movement in this way.