9

In 2001 M.I.T. Press published a volume titled Parasitic Gaps, edited by Peter W. Cullicover and Paul M. Postal. Its preface begins as follows:

Parasitic gaps (P-gaps) represented by the underlined gaps subscripted p in example (1), have been an intensely studied subdomain of natural language syntax for almost twenty years.

(1) (a) It was Ernest who pictures of _____________p tended to depress _______________.

     (b) No matter which candidate the reporter criticized ________________ immediately after meeting ______________p, he won't vote for you.

Can this be translated into language that we unordained laypersons can understand? In both cases it is easy to imagine some word or phrase that it would make sense to put there the p-subscripted blank is, and both cases the other blank seems to be something that should be left blank. The authors seem to think they are telling the reader what a "parasitic gap" is, but I can't make anything of it.

Michael Hardy
  • 665
  • 3
  • 8
  • Whatever else, it looks like you're quite right to question those examples. My own belief is that in language laypersons understand, both are meaningless in general language, which necessarily makes them useless as examples.

    By the way, what's this: in 2001 M.I.T. Press realised that parasitic gaps had been intensely studied for almost 20 years? How could anyone think that meant other than that p-gaps were pretty-much a brand-new thing and therefore worthy not of less but of much more careful explanation?

     – Robbie Goodwin Jun 15 '21 at 22:54
  • @RobbieGoodwin : I doubt that M.I.T. Press realized anything. Rather, M.I.T. Press decided to publish a book by persons who had realized something. – Michael Hardy Jun 16 '21 at 03:26
  • Thanks for the put-down and how are those different, in practical terms? – Robbie Goodwin Jun 16 '21 at 20:05
  • @RobbieGoodwin : I don't know why you think anything here was a put-down. – Michael Hardy Jun 17 '21 at 05:29

2 Answers2

16

A parasitic gap is a gap that is reliant on the presence of another gap; it is "parasitic" on the other gap. Gaps in general mark positions in the syntactic structure that are empty because what fills them appears elsewhere in the sentence, e.g.

(1) Frank rejected the first explanation without even considering it.

(2a) Which explanation did Frank reject ___ without even considering it?

Sentence (1) is a normal statement containing no gaps. The question (2a), in contrast, contains a gap, which is marked by "___" and filled in the relevant sense by the fronted wh-expression which explanation. This gap is normal, i.e. not parasitic in any way. We can now add a second gap to the question, though:

(2b) Which explanation did Frank reject ___ without even considering ___?

In this case, the second gap is parasitic, i.e. reliant, on the first gap, for it cannot appear without the first gap. Compare:

(2c) *Which explanation did Frank reject it without even considering ___?

(3a) *Frank rejected the first explanation without even considering ___.

(3b) Frank rejected the first explanation without even considering it.

The Wikipedia article on parasitic gaps provides many more examples with accessible discussion. Have a look:

https://en.wikipedia.org/wiki/Parasitic_gap

Tim Osborne
  • 5,747
  • 1
  • 17
  • 40
  • 1
    3a is not ungrammatical, it just means something slightly different (‘without consideration in general, without thinking, unhesitantly’, rather than ‘without considering the explanation’). – Janus Bahs Jacquet Jun 12 '21 at 08:46
  • 2
    @JanusBahsJacquet: It's standard, in this sort of context, to describe an utterance as "ungrammatical" if it's ungrammatical under the intended interpretation, even if there's some other interpretation under which it would be grammatical. – ruakh Jun 13 '21 at 04:10
  • 1
    (3a') *@JanusBahsJacquet dismissed this answer without even thinking about ___. – TonyK Jun 13 '21 at 15:59
6

The key property of parasitic gaps is that p here represents the same constituent, such that one of the gaps is "parasitic on" (i.e. coreferential with) the other.

For example, in 1b), both the gaps are coreferential with the word candidate. You can understand the semantics as:

there exists no candidate, such that if the reporter criticized them immediately after meeting them, then it would cause the reporter to vote for you.

Because "them" occurs twice, referring to the same candidate, any syntactic phenomenon which extracts it will leave two gaps, and the second gap is said to be parasitic on the first.

Incidentally, Combinatory Categorial Grammar (CCG) admits a neat computational analysis of parasitic gaps using the S (substitution) combinator -- more details are found in Steedman (2000), but here is an example from my dissertation (note that "cake" is both the thing that you can eat, as well as the thing you don't have to bake -- a parasitic gap!)

parasitic gap example

1: Steedman, M. "The Syntactic Process" (2000). https://mitpress.mit.edu/books/syntactic-process

jogloran
  • 5,145
  • 28
  • 47