Strict Polarity

More on this in the refgram
Syntax: Polarity_phrases

The Strict Polarity proposal limits where bu and words like it (polarity items, Σ) can occur.

The problem

What does bu marao jí ("I don't dance") mean?

Officially, bu "can go anywhere" and "simply inserts a ¬ in place." Hoemaı's intent is for bu marao to claim an event of non-dancing.

The refgram says that polarity items "usually attach to 𝑣Ps or TPs."

But if the 𝑣P marao jí denotes λe. marao(e) ∧ agent(e) = jí, then bu marao jí means λe. ¬(marao(e) ∧ agent(e) = jí). Rather than "I do some non-dancing", this claims "Something happens, but it's not 'me dancing'", which is a strange meaning. It gives us almost no information about e, and it isn't really anything like what we mean by "I don't dance".

So, negating a 𝑣P (by placing bu before it) is apparently a bit weird. How do we get bu marao jí to mean something sensible?

Other sensible negation scopes

Let's suppose for a moment that we have a mechanism to negate verbs: ⟦bụmarao⟧ = λe. bumarao(e), an event of non-dancing, which we can then say jí participates in.

Negating a TP has different, but also intuitive semantics: Bu mala marao jí means "I've never danced" (a denial of Mala marao jí).

Maybe these are enough.

Strict Polarity

We can limit negation to these two cases: bu- negates verbs and bu negates TP.

1. bụmarao jí claims an event of my non-dancing.
2. bu marao jí is actually bu (tuom) (tam) marao jí, and it means "there is no event of me dancing at that time."

There is always a hierarchy like (ΣP) TP AspP 𝑣P (Σ-)V instead of Σ heads showing up anywhere.

Sentences like Pu bu chum hao jí or Pu chum bu hao jí are now ungrammatical.

Conjunctions like rú now join ΣP instead of TP.

So how does bu- work?

There are more questions than answers here.

If we say bu- is just ¬, then the verb bụmarao denotes λe. ¬marao(e), which is still pretty weird. We want to say e is an event of not-dancing: an event of refraining from dance. But λe. ¬marao(e) only says what e isn't, which isn't specific enough.

So, maybe bu- is not simply ¬. It's a bit unclear how to get good semantics for "not-dancing".

1. One idea is that bụmarao(e) means: e is an event with which no marao-event is concomitant.
2. Another idea is that bụmarao-ing is a verb whose truth conditions are inverted from marao.
3. A promising idea from semantics research is that bu- makes "negative events".

What about transitive verbs? It seems ridiculous to say that we have successfully bụchuq-ed for a week simply because there exists a cookie we haven't eaten all week. So maybe it's sensible for bu- to turn verbs intransitive: bụchoa is to keep one's mouth shut (as to not-speak is to not speak anything), bụchuq is to fast (as to not-eat is to not eat anything), and so on.

See also

• Negation