Comments on Kayne and Moro 2022

Kayne and Moro 2022 observe that zero in English can license weak NPIs:

1.

a. I have zero interest in doing anything right now (= (3))

b. Zero students have ever said anything like that. (=(6))

I agree with these judgments. Kayne and Moro account for these facts in terms of a silent NOT (capitalization indicates that NOT is syntactically present, but not pronounced):

2.

I have NOT zero interest in doing anything.

The underlying assumption is: “…NPI licensing is in all languages invariably due to the presence of a negative morpheme…” (pg. 2) Of course, to maintain such a strong position, Kayne and Moro will be obliged to explain why one gets weak NPIs in the restriction of universal quantifiers, amongst many other contexts: Everybody who knows any physics takes my class. I do not pursue this issue here.

Kayne and Moro give additional evidence for a silent NOT based on combien extraction, the distribution of negative ne and negative inversion. I will not comment on these additional facts here, but focus on the licensing of NPIs.

I propose to reanalyze the basic NPI facts in (1) not in terms of a null negation NEG, but in terms of null ONLY or perhaps null EXACTLY. On this approach, the structure of (3a) is (3b):

3.

a. Zero students knew any physics.

b. EXACTLY/ONLY zero students knew any physics.

Now both only and exactly license weak NPIs:

4.

a. Only John knows any physics.

b. *John knows any physics.

5.

a. Exactly three students know any physics.

b. *Three students know any physics.

The difference between (3a) and (5b) is especially telling. Normally numerals (such as three) modifying nouns do not license NPIs, so it is curious that zero (a kind of numeral) does.

Why is this approach superior to that of Kayne and Moro 2022. Kayne and Moro’s approach seems to force them to say that zero means something like “some” in order to get the interpretation right. They admit this much (pg. 3): “The question now arises as to the status of zero itself, in the context of NOT. We might take it to be a (rather special) subtype of numeral. Or it might be closer to some, any, several, a few, or a number of.”

NEG combining with some should yield a negative sentence (personal communication, Paul Postal) for the purposes of the Klima tests. But zero followed by a positive tag seems unacceptable:

(6)

a. Nobody came to the party, did they?

b. Not a single person came to the party, did they 

c. *Zero people came to the party, did they?

However, only or exactly followed by a positive tag is also unacceptable:

(7)

a. *Only John came to the party, did he?

b. *Only three people came to the party, did they?

c. *Exactly three people came to the party, did they?

Similarly, there is a clear difference in slifting:

(8)

a. Nobody came to the party, I don’t think.

b. Not a single person came to the party, I don’t think. 

c. *Zero people came to the party, I don’t think.

(9)

a. *Only John came to the party, I don’t think.

b. *Only three people came to the party, I don’t think.

c. *Exactly three people came to the party, I don’t think.

In effect, I am arguing that zero does combine with negation, but not in the same sense as Kayne and Moro 2022. Only and exactly have a negative component, as shown by paraphrases with overt negatives plus exceptives, etc.

(10)

a. Only John came to the party.

“Nobody but John came to the party.”

b. Exactly three people were there.

“Three people, no more and no less, were there.”

Whether or not only and exactly involve a syntactic NEG morpheme is different question for a different day.

One avenue to explore is whether other numerals allow null ONLY and EXACTLY in the same contexts that zero does.  

References:

Kayne, Richard and Andrea Moro. 2020. A Note on Zero and Silent Negation. Ms., New York University and University School for Advanced Studies IUSS Pavia.