1

Are there any phrases were a conventional grammar gives multiple syntactic analysis for the same phrase (without changing the meaning) or even fails to give a consistent structure?

I thought about this because I was wondering whether grammars were totally consistent or they presented some ambiguities.

Would word classification be a similar case?

  • 3
    Well, for starts, as Abney has demonstrated, virtually every written English sentence is multiply ambiguous syntactically. Writing doesn't include intonation, rhythm, voice quality, gaze direction, facial expressions, gestures, or other parts of language that provide the context to disambiguate the syntax. And of course there are always plenty of non-compositional phrases like kick the bucket or the wrong number. – jlawler Jul 3 '14 at 18:03
  • 1
    According to what theory of syntax? Which model of grammar? You need to be a lot more specific. – curiousdannii Jul 3 '14 at 22:10
  • The most pervasive and universal trait among natural languages may be ambiguity. – hippietrail Jul 4 '14 at 2:28
  • 2
    @hippietrail and it's a feature, not a bug! – Gaston Ümlaut Jul 4 '14 at 5:12
  • @jlawler But wouldn't the syntax in those cases be unique for a fixed meaning? – jinawee Jul 4 '14 at 7:41
1

If ambiguity is what the question is most concerned with, I think the short answer to the question is no: distinct structural analyses of an utterance almost always point to distinct meanings.

There are, however, numerous sources of ambiguity, syntax being one and lexical meaning being another. If the grammar does not have a way of acknowledging the cues that disambiguate the utterance at hand, an incorrect structural analysis is easily possible. Consider the following sentence in this regard:

 We like the baby in the crib. 

This sentence might mean that there are two babies, one in the crib and one on the floor, and we like the one in the crib, not the one on the floor, or it could mean that we like to keep the baby in the crib as opposed to, say, on the floor. The difference in meaning is captured in the syntax in terms of whether in the crib is a dependent of baby or of like. A second example:

 a Peruvian silver tray

This noun phrase might mean that the tray is made of Peruvian silver, or it could mean that they tray is Peruvian, whereby the silver may or may not be Peruvian. The ambiguity is again captured in the syntax in terms of whether Peruvian is a dependent of silver or of tray.

Conventional grammar is going to fail to produce a plausible structural analysis of these cases if it does not have a means of disambiguating the meaning. The cues that are important for disambiguation are usually present in context, but if the grammar does not have a way to register these cues, the danger of producing the wrong syntactic analysis is great.

In the case of a Peruvian silver tray, the disambiguation cue comes in the form of intonation. If the phrasal accent is on Peruvian, i.e. a PeRUvian silver tray, the tray is made of Peruvian silver, but if the accent is on silver, i.e. a Peruvian SILver tray, the tray is Peruvian. Concerning the first example, i.e. We like the baby in the crib, I am not aware of an intonational cue that disambiguates the utterance. I assume, rather, that the general meaning context would disambiguate the utterance, e.g. whether or not there is more than one baby in the discourse situation.

So again, if the grammar being used does not have a means of acknowledging these cues, it will often produce an incorrect syntactic analysis.

A similar situation obtains for lexical ambiguity, e.g.

 I am blue today. 

This sentence could mean that I am wearing blue today, or more likely, it means that I feel sad today. This ambiguity resides completely with the meaning of blue; it is therefore not in the syntactic structure, since both meanings have the same syntax; blue is an adjective in both cases. The disambiguation cue is present in the context of utterance. If I am standing in front of you wearing blue, then it is clear which meaning is intended. Syntax, however, has little to do with such cases.

All that said, there are in fact certain cases in which two underlying syntactic structures are assumed for one and the same unambiguous phrase. This has been proposed for certain noun phrases, e.g. the old fish with stripes. Such noun phrases behave in a unique way with respect to diagnostics for constituent structure, and this has motivated some to propose that there are in fact two distinct underlying structures for such prases. In my view however, that sort of approach is a cop out. One might as well throw up one's hands and declare that one has no idea what is behind the confusing results delivered by the diagonostics.

| improve this answer | |
  • I'd think that if the phrasal accent is on "Peruvian" (a PeRUvian silver tray), it would mean "a silver tray from Peru" whereas if it's on "silver" (a Peruvian SILver tray), it'd mean "a tray made of Peruvian silver". – Morphosyntax Jul 9 '14 at 4:18
  • I think my interpretation in the answer is correct, but I acknowledge I may have got it wrong. The example is actually from Allerton (1979: 116). Allerton claims that the phrase is actually three-way ambiguous. Perhaps he's right, which would mean there is a third reading that could be the source of my/our lack of clarity in analysis. – Tim Osborne Jul 9 '14 at 7:12

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.