Discussion Groups / English Usage / January 2004

I'm not making myself very clear, it seems; and I'm beginning to think
I may not be able to. I'll have another go, but forgive me if it still
doesn't get through.

"Last" is here an adjective; "year" is a noun; the two together,
though, make up an adverbial phrase of time when used in sentences
such as "I saw him last year". This phrase can answer the question
"When?"

On the other hand, "last year" can also be a noun phrase when used in
sentences such as "I enjoyed last year". In this use, the phrase can
answer the question "What?"

In our sentence, "last year" starts as a noun phrase: "What will you
mention?" "I shall mention last year." It then turns into an adverbial
phrase: "When did he talk sense?" "He talked sense last year".
Interpreting it as an adverb makes the when-clause logical in a way
which would not be possible with a sentence such as "*I shall mention
Australia, when he talked sense". The relative adverb "when" must, in
formal use, have an adverbial time-expression as antecedent.

This dual function can be seen as an anomaly: "I shall mention last
year" makes sense, as does "Last year, when he made sense, I paid
attention to him." But it's clear that in these examples the phrases
don't perform the same function: in the first, it's a noun; while in
the second, it's an adverb. In our sentence, though, the same phrase
seems to have to perform both functions, since, as I said, a classical
adverb or an adverbial clause cannot modify a noun.

I couldn't agree with Rolleston when he implied that an adverb could
modify only a verb: an adverb may modify a verb, adjective, or another
adverb: it was on that ground that I was challenging his analysis.

My personal view, for what it's worth, is that the only real problem
here is that our terminology is faulty! "When" is a more flexible word
than a traditional view might allow.

Mike.

But "last year" appears in two contexts in the same sentence: in one
of these it's clearly a noun phrase, while in the other it's as
clearly adverbial -- because an adverbial clause cannot modify a noun.
We can't choose which is of "higher rank", so we may have to decide
that it is being used as both. Conversational English is so full of
this kind of thing that we are forced either to the uncontroversial
conclusion that the terminology is inadequate, or to build an edifice
of rules explaining them away by purely hypothetical "colloquial
ellipsis" as we could in this case.

Or we could try to rehabilitate the terminology by ruling that under
some circumstances in English an adverb may modify a noun; or that
under some circumstances in English "when" may be a relative pronoun:
but what would be the point of that? These terms have pretty
well-established meanings for the languages for which they were
devised, and I'd prefer not to tinker.

Mike.

Aha! At least something agrees with my Latin grammar.

This is all very interesting. I wonder, could you recommend a
good English grammar textbook of the sort aimed at, say, 1st
year undergraduate linguists? I'm out of my depth here.

I think there is a problem with these commas for non-restrictive
relative clauses. To my mind they've been forced into a role
for which they are not suited. The relative pronouns that begin
these clauses seem quite strongly drawn to their antecedents. So
the comma seems inappropriate. In an ideal world there would always
be alternative pronouns to use to indicate the kind of relative
clause that would follow. And the comma isn't much use when something
surrounded by commas gets wedged between the antecedent and the pronoun.
Then you can't tell whether the comma would have been there if the
inserted bit was missing.

R.

I wasn't claiming that the third sentence contains a relative clause.
I was just applying an apposition diagnostic to the first sentence.
The fact that the diagnostic puts relative clauses like the one in the
first sentence in the category of appositions suggest that the
diagnostic is very weak (prone to false positives) or that relative
clauses and appositions are actually quite closely related.

My point was that we can't conclude that "last year when he talked
sense" contains a relative clause on the basis of Rolleston's
arguments alone, because there is another type of construction
(appositions) that these arguments don't rule out.

I'll have a read.

Yes, you're quite right.

I wasn't claiming that the third sentence contains a relative clause.
I was just applying an apposition diagnostic to the first sentence.
The fact that the diagnostic puts relative clauses like the one in the
first sentence in the category of appositions suggest that the
diagnostic is very weak (prone to false positives) or that relative
clauses and appositions are actually quite closely related.

My point was that we can't conclude that "last year when he talked
sense" contains a relative clause on the basis of Rolleston's
arguments alone, because there is another type of construction
(appositions) that these arguments don't rule out.

I'll have a read.

Yes, you're quite right.

Yes, that's what it means. The point is that relative clauses are not
the only kinds of things that can modify noun phrases so saying x
modifies a NP, doesn't mean that x is necessarily a relative clause.

Good point. I would feel uncomfortable calling a year an event too.
Perhaps it is better to say that these time phrases can be used with
any noun that labels a thing that can be positioned in time. This
would encompass things like events and years.

Note that the first sentence is ambiguous. The speaker might have
remembered a day that fell in June, or may have remembered, in June, a
day that fell possibly in some other month. The second sentence is not
ambiguous - the remembering happened in June.

No, but the ambiguity in my example can be used to justify the view
that the 'adverbial' can attach to either a verb or a noun - that is,
it can justify the view that there are two different syntactic
structures corresponding to the two different interpretations of
examples like "I read the letter in the kitchen".

Your use of contrasting tenses to construct your example was a good
move in my opinion. You've given us a compelling reason to believe
that, in your example, "when he talked sense" modifies "last year"
rather than the verb. If the tenses were the same, it would have been
ambiguous between a structure in which the 'when' clause attaches to
the verb and a structure in which it attaches to the noun:

"I mentioned last year when he talked sense."
Ambiguous between at least these two meanings:
1.  At the point when he talked sense, I mentioned last year.
2.  I mentioned last year, which was when he talked sense.

You've spared us this ambiguity, but the NP "last year when he talked
sense" is still ambiguous between the reading in (2) and this:
3.  the point during last year when he talked sense.

As in (2), the 'when' clause is modifying "last year", but this time
it is restricting the referent, narrowing it down to a more specific
period of time. So the 'when' clause can either narrow down the period
of time you are referring to (as in meaning 3) or provide incidental
information about "last year" (as in meaning 2). It's quite a subtle
kind of ambiguity, but it's there.

Your question, in sharp focus, is then "What is the analysis when read
with a non-restrictive meaning like 2?"

You could approach this by arguing for the _best_ analysis - the one
that is most general and covers the most data. The three analyses
we've considered for the 'when' clause are:

1. It's a non-restrictive relative clause.
2. It's an NP-internal time phrase ('abverbial')
3. It's an apposition.

And we could compare some others if we can think of any. And then
there's the question of whether these analyses can be unified. Perhaps
they are actually different manifestations of the same underlying
phenomena.

H.
 
---
Like-minds don't notice shared mistakes. Talk to someone else.