Re: Which part of speech?

David J. Peterson <dedalvs@...> wrote:
>Damian wrote:
><<
>> So then how do you parse "last night" as an adverb?  Wouldn't
>> it just be simpler (in the Occam's Razor sense) to assign a
>> zero-derived case to every noun naming a day?
> >>
>
>Simpler in what sense?  It would actually seem needlessly complicated
>to me.
You generate fewer lexical entries (several dozen fewer, at the least) and
don't need a category of adjectives that modify adverbs with the null case /
null preposition method.

Or you could have a null morpheme to change noun phrases into adverbial
phrases. However, since there's already a way to make noun phrases modify
verb phrases, why make a new one? Only if the old one can't account for data.


>I'd say a more accurate way to characterize what Chris wrote above is
>that, in English, these are *not* nouns (referring to "today",
>"yesterday",
>etc.)--or, at least, they're not nouns when they're adverbs.  They
>can be
>nouns, e.g.,
Sure, if by "more accurate" you mean "completely contradictory". You're
obviously building a noun phrase--you can modify it with adjectives rather
than adverbs, so it's not an adverb phrase, adjective phrase, or verb
phrase. It must be a noun phrase.

Around the noun phrase, you must construct some sort of framework to allow
it to modify a verb phrase. I favor using a null preposition; Damian favors
using a null case marker. You could also use a null adverb marker, but since
we can't normally turn nouns directly into adverbs (*devilly, but attested
devilishly), I'd prefer a different explanation.

But if you used an adverb rather than a noun in every such construction,
then you'd need an adverb "next_week", an adverb "the_week_after_next",
"three_weeks_from_now"....Or you'd have to come up with a method for
constructing adverb phrases that look just like noun phrases, have
adjectives rather than adverbs as modifiers, and use quantifiers and
determiners.

Since it's so far-fetched that one could allow a radically different
construction in a certain class of temporal adverbs that mimics nominal
structures, I'd vastly prefer to say that these are in fact noun phrases.


>(b) I ate a cake yesterday.
>(c) Yesterday I ate a cake.
>
>A null preposition seems like a good idea, since you can replace
>"yesterday"
>in (b) with any prepositional phrase that makes sense.  However, in (c),
>I get the distinct impression that you can't comfortably.  Compare
>the two
>below:
>
>(d) I ate a cake on the lawn.  [eater is on the lawn]
>(e) On the lawn I ate a cake.
>
>Both are grammatical.  However, at least to me (and others can judge if
>this is accurate), I get the sense that (b) and (c) are equally
>neutral--one
>is liable to utter either in isolation.  For the latter pair, though,
>(d) is neutral,
>and (e) is not.  There must be *some* kind of prompt to front the PP in
>(e), whereas no prompting is needed to front the adverb in (c).
>Also, if
>you replace the locative PP with a manner PP, it gets worse:
>
>(f) I cut the cake with a knife.  [instrumental reading]
>(g) With a knife I cut the cake.
>
>Again, both are grammatical, but I get the sense that (f) is neutral and
>(g) definitely is not.
I work mainly with minimalism, so my minimalist instincts kick in here to
suggest a solution.

Temporal phrases are TP adjuncts (TP = tense phrase = S). Locative and
instrumental phrases are vP adjuncts (vP = verb phrase). Adjunction has a
loose notion of directionality, so you can say "I went to the store
yesterday" and "Yesterday I went to the store" without movement. Either way,
"yesterday" appears on the edge of TP.

On the other hand, if you want the phrase "with a knife" to appear first in
a sentence, you have to move it at least to the edge of TP. It's an example
of topicalization, not a stylistic change in directionality of an adjunct.


>So while positing a null preposition does seem "attractive", it doesn't
>seem to account for the data.  If there is no null preposition, then one
>is forced to try to come up with an explanation for why certain nouns
>get automatic case, but only in certain situations (e.g., in (a), the
>preposition "for" would be assigning case to "tomorrow", under a
>case-assigning theory).  Either that, or things like "tomorrow" don't
>need to have case, and, hence, are adverbs.
It doesn't account for all the data, perhaps. You won't find one account of
one problem that accounts for all the data, only for a subset of the
data--and that's fine. We have multiple theories within each theoretical
framework for just that purpose.

Right now, I'm working on a description of multiple WH questions in English.
This description includes two theories to account for the sentence "Where
did John park his car?" and one more to account for "Who sent a letter to
who?" And I'm not sure how I'll account for "how many / how much" at this
point. So if we don't achieve a full explanation of temporal phrases with
one statement, I'll not be surprised.


>Or, one can simply say that the English data doesn't need to be
>accounted for, since it doesn't fit the theory.  That's a very popular
>option.
A running joke among linguists--if the data doesn't fit your theory, the
data's probably wrong.


>See below for a paper contrasting -ly adverbs and temporal/locative
>adverbs (in HTML form):
>
>http://scholar.google.com/scholar?
>hl=en&lr=&c2coff=1&safe=off&q=cache:jvJaBsbEibUJ:www.sakura.cc.tsukuba.a
>c.jp/~hidekazu/Paper/ly-adverb.pdf+temporal+adverbs+as+nouns
Note that that resource listed the phrase "at the store" as an adverb. You
won't find many linguists that will agree. It's a prepositional phrase. You
have a noun, a determiner, and a preposition; thus, there's a noun phrase
and a prepositional phrase (and according to minimalism, at least, a
determiner phrase).

So if your definition of "adverb" includes prepositional phrases, then yes,
the string "yesterday" can be an adverb, but only if it's a proper noun (to
account for the lack of a determiner) with a null preposition. Most
linguists list adverbs as words, not phrases, and distinguish prepositions
from adverbs. At least the ones I've encountered; anyone using a Chomskyan
approach to syntax does.

What syntactic framework are you using? Or rather, what's being used at the
University of Tsukuba, from which you got that paper?

-Chris Wright

Replies