[Grammar] complements and the one-substitution test

[raymondaliasapollyon]

Hi,

Formal syntax literature has discussed a test known as the one-substitution test, whereby "one" is said to contain a complement. Robert D. Borsley in his Syntactic Theory: A Unified Approach gives the following examples:

a. This student of linguistics from Korea is more dedicated than that one.
b. This student of linguistics from Korea is more dedicated than that one from China.
c. * This student of linguistics from Korea is more dedicated than that one of psychology from China.

Since one contains a complement, the one in b must be understood as "student of linguistics." And c is ungrammatical precisely because one is superfluously followed by another complement, i.e., of psychology.

Having said that, I'd like to draw your attention to a variant of a sentence I mentioned in a previous thread:

d. You know about addictions to drugs and cigarettes, but here's one you may not be cognizant of: fitness addiction.

The one-substitution test would suggest that one would contain a complement and be equal to "addiction to drugs and cigarettes" here, If so, the sentence should be ungrammatical. (It would be equal to "You know about addictions to drugs and cigarettes, but here's an addiction to drugs and cigarettes you may not be cognizant of: fitness addiction.")

Do you find sentence d ungrammatical?

 

[Phaedrus]

d. You know about addictions to drugs and cigarettes, but here's one you may not be cognizant of: fitness addiction.

The one-substitution test would suggest that one would contain a complement and be equal to "addiction to drugs and cigarettes" here, If so, the sentence should be ungrammatical.

What proof do you have that "to drugs and cigarettes" is not functioning as an adjunct instead of as a complement? We could imagine "to drugs and cigarettes" even being added as an afterthought, in which case it can hardly be claimed to be requisite for completion of meaning. How, then, are you going to insist that the to-PP is a complement? In the sentence below, "one" is obviously equal to "one addiction." If the to-PP can be a nonrestrictive modifier/adjunct here, why can't it be a restrictive one in (d)?

You know about (many different) addictions—to drugs and cigarettes and whatnot—but here's one you may not be cognizant of: fitness addiction.

 

[raymondaliasapollyon]

What proof do you have that "to drugs and cigarettes" is not functioning as an adjunct instead of as a complement? We could imagine "to drugs and cigarettes" even being added as an afterthought, in which case it can hardly be claimed to be requisite for completion of meaning. How, then, are you going to insist that the to-PP is a complement? In the sentence below, "one" is obviously equal to "one addiction." If the to-PP can be a nonrestrictive modifier/adjunct here, why can't it be a restrictive one in (d)?

You know about (many different) addictions—to drugs and cigarettes and whatnot—but here's one you may not be cognizant of: fitness addiction.

Adjuncts are added rather freely, but there is a sense in which the to-PP is in a special relation to "addiction." If an addiction exists, it must be to something. The to-PP thus completes the meaning of "addiction."

Consider "on the desk" in "The book on the desk is blue." The PP on the desk can be added to almost any NPs as a modifier, hence an adjunct.

Do you know of any independent tests for adjuncthood vs. complementhood?
Can't complements occur as afterthoughts?

 

[PaulMatthews]

Adjuncts are added rather freely, but there is a sense in which the to-PP is in a special relation to "addiction." If an addiction exists, it must be to something. The to-PP thus completes the meaning of "addiction."

Consider "on the desk" in "The book on the desk is blue." The PP on the desk can be added to almost any NPs as a modifier, hence an adjunct.

Do you know of any independent tests for adjuncthood vs. complementhood?
Can't complements occur as afterthoughts?

If memory serves, the one-substitution test uses the pro-nominal "one", not the determinative "one" used in your example You know about addictions to drugs and cigarettes, but here's one you may not be cognizant of: fitness addiction.

Incidentally, I think it's preferable to restrict the term 'adjunct' to modifiers in clause structure. Thus the PP "on the desk" is simply 'modifier' of "book".

 

[raymondaliasapollyon]

What's a determinative "one"?
Never heard about it in the literature.

 

[PaulMatthews]

What's a determinative "one"?
Never heard about it in the literature.

Pro-nominal "one" is a common noun that functions as head in NP structure. Determinative "one", as its name implies, functions as determiner (or modifier), either with a following head or in fusion with the head:

[1] This knife is blunt: have you got [a sharper one]? [pro-nominal "one" as head]

[2] This brush won't do: I want [one with a handle]. [determinative in fused determiner-head construction]

The difference is that if we expand to a non-reduced NP we simply replace pro-nominal "one" by a new, non-reduced head, i.e. "a sharper knife", whereas the determinative "one" in [2] is replaced by a sequence of determiner + head , i.e. "a brush with a handle".

 

[raymondaliasapollyon]

Is the following okay?

John is a professor of linguistics, but we need one of physics.

The one here is a determinative one.

 

[PaulMatthews]

Is the following okay?

John is a professor of linguistics, but we need one of physics.

The one here is a determinative one.

Yes, "one" is understood as "a professor" (determiner+head).

 

[Phaedrus]

[. . .] there is a sense in which the to-PP is in a special relation to "addiction." If an addiction exists, it must be to something.

Nice point.

The to-PP thus completes the meaning of "addiction."

I'll tentatively concede that point. Unlike your of-PPs, however, the to-PP may be said to answer a question about the type of addiction.

A: He has a couple of bad addictions?
B: I'm sorry to hear that. What type(s) of addictions does he have?
A: He's addicted to drugs and cigarettes. / He has an addiction to drugs and (an addiction to) cigarettes.

That might suggest that the PP is a modifier. The same question can be answered by adjectives in other cases. We wouldn't want to call modifying adjectives complements.

A: What type of rice do you have?
B: Brown rice.

But your of-PPs don't work well for questions about type.

A: He's a student.
B: What type of student is he?
A: ?? He's a student of history. (Vs.: He's a hard-working student.)

Do you know of any independent tests for adjuncthood vs. complementhood?

I've been wanting to be able to answer "Yes" to that question for quite some time. Unfortunately, I'm not there yet.

 

[Phaedrus]

Pro-nominal "one" is a common noun that functions as head in NP structure. Determinative "one", as its name implies, functions as determiner (or modifier), either with a following head or in fusion with the head:

[1] This knife is blunt: have you got [a sharper one]? [pro-nominal "one" as head]

[2] This brush won't do: I want [one with a handle]. [determinative in fused determiner-head construction]

The difference is that if we expand to a non-reduced NP we simply replace pro-nominal "one" by a new, non-reduced head, i.e. "a sharper knife", whereas the determinative "one" in [2] is replaced by a sequence of determiner + head , i.e. "a brush with a handle".

Excellent distinction. I think that there is, in addition to those two syntactic categories of one, a third one (tee hee)—namely, the category of cardinal numeral.

I believe it is this third category of "one" that is found in the example in Post #2 (and maybe the OP), as may be demonstrated by the use of other numerals:

You know about many different addictions, but here are two you may not be cognizant of: fitness addiction and syntax addiction.

Two introduces NP ellipsis there, of course (two = two addictions). One does the same thing in post #2 and maybe in the OP as well.

 

[PaulMatthews]

Excellent distinction. I think that there is, in addition to those two syntactic categories of one, a third one (tee hee)—namely, the category of cardinal numeral.

I believe it is this third category of "one" that is found in the example in Post #2 (and maybe the OP), as may be demonstrated by the use of other numerals:

You know about many different addictions, but here are two you may not be cognizant of: fitness addiction and syntax addiction.

Two introduces NP ellipsis there, of course (two = two addictions). One does the same thing in post #2 and maybe in the OP as well.

Determinative "one" is a cardinal numeral, though it isn’t necessarily interpreted that way. We have to rely on the antecedent for the relevant interpretation, cf:

[1] We need three keys, but we’ve [only one].
[2] I’ve come out without a pen: can you lend me [one]?

In [1] "only one" is equivalent to "only one key": here it does contrast with larger cardinal numerals. In [2] on the other hand, "one" is equivalent to "a pen": here "one" is not in competition with cardinal numerals, but behaves like a stressed form of the indefinite article.

 

[Phaedrus]

Is the following okay?

John is a professor of linguistics, but we need one of physics.

The one here is a determinative one.

That example sounds pretty bad to me, though I wouldn't call it ungrammatical. I'd prefer:

John is a professor of linguistics; we need one in the field of physics.

That said, the following sounds perfectly good to me: That is a cup of salt; we need one of sugar.