Semantics - Negation

Where does a negation go in an English sentence?

After the tense (b/w tense + VP)

Scope

Whats in the domain of the operator

Ex. in “she will not walk”, “not walk” is negated (but any can be negated depending on how its said)

Are negations hard to cog process?

Yes - more effortful + takes longer

How do we process negated sentences when we are getting 1 word at a time and the negation is later on

Negation

As a logic operator, negation returns the opposite truth value as its input

How do we update the representation of the sentence’s meaning?

Build meaning representation of proposition + then add the negation after

In steps, how we compute meaning of “the butterfly is not above the hummingbird”?

1. Compute “the butterfly is above the hummingbird”

2. Negate the truth value

Compute core proposition first then negation after

A is not above B
not (A is above B)

What can we use to find out how negation affects lexical expectation?

N400; Bc the N400 should be sensitive to truth value since its unexpected

2 Theories for explaining how negation affects lexical expectation?

1. 2 Step Processing

2. Incremental Processing

2 Step Processing

Expectation at word n is a function of the meaning of the core proposition only; Negation is computed after we have parsed the entire core proposition (negation updated only at the end)

2 Theories for explaining how negation affects lexical expectation?

Incremental Processing

Expectation at word n is a function of the meaning of the core proposition - but that expectation is changed by the negation right away; Expectations instantly updated

Ex. A robin… (thinking abt possible continuations like a bird that can fly, eats words etc)

Ex. A robin is… (restricts prediction a bit; a bird sitting in the tree)

Ex. A robin is not… (previous ones no longer work! its a lot to compute what a robin isnt)

Ex. A robin is not a bird (not expected; will get neg expectation for this)

Fischler et al Study

Predicted that since N400 is a function of lexical/semantic expectation, negation should REVERSE semantic expectation (aka negation should dramatically affect N400)

Used 2 statements with a fixed truth value; Maybe false sentences are surprising + trigger an N40

Ex. A robin is a… (bird → no n400) (affirmative)
Ex. A robin is not a… (bird → n400 bc false) (neg)

Results → Got N400 for false statement in ex 1 AND true statement in ex 2; Shows maybe N400 has to do with priming

A robin is a bird → No N400

A robin is NOT a bird → No N400

A robin is a hammer → N400

A robin is NOT a hammer → N400

Affirmative F sentence → N400

Neg T sentence → N400

Fischler et al Conclusion

Negation does not affect N400; If prior context modulates lexical expectation, negation should REVERSE expectation

Does cloze probability correlate with N400?

Yes

Is N400 sensitive to context?

Yes; Could mean that sentences that are surprising when they are NOT natural

Nieuwland & Kuperberg 2008 Study

Predicted would get N400 when a response is unnatural (based on Fischler et al)

Made sentences with a high construct that inverts expectation

With proper equipment, scuba diving is very… (safe)

With proper equipment, scuba diving isnt very… (dangerous)

Got N400 for:

With proper equipment, scuba diving is very dangerous

With proper equipment, scuba diving is not very dangerous

Have a case where negation seems to matter and one where it doesnt seem to matter; Negation matters DEPENDING ON CONTEXT (negation takes time to immediately override expectation given context; meaning we associate proper equip with safe, and anything else would be surprising)

Haase, Spychalska & Werning (2019) Hypothesis

Maybe we see N400 due to semantic relation/mismatch between subject + predicate

Ex. robin…bird (no N400)

Ex. robin…hammer (N400)

Are sentences expected to be informative?

Yes; But we already know things like a robin is a board

Palaz, Rhodes & Hestvik (2020) Study

Wondered if negation is processed incrementally when sentences are informative; And what if you are leaning new words?

Had a novel word learning paradigm where participants guessed the meaning of novel words from the context of a sentence, then rate a follow up sentence as plausible or implausible (EEG measured when rating plausibility)

Found that we are NEGATION BLIND

Ex. I saw a trante with exotic gills at National Aquarium → trante is a fish (no N400)

Ex. I used a huge sloken to bang nail in wall → A sloken is a fish (N400)

Ex. I saw a trante with exotic gills at the National Aquarium → A trante is NOT a fish (no n400)

Ex. I used a huge sloken to bang the nail in the wall → A sloken is NOT a fish (N400)

Naturalness

Sentences are unnatural bc they are uninformative; But informative when you are learning new words

In the real world, we expect utterances to be informative; A hammer is not a bird is not informative, so it’s hard for us to know what to expect

Do we see negation blind results in environmental sound priming?

Yes