A Role for Processing in Basic Language Cognition: Commentary on Hulstijn (2024)

Abstract

This brief commentary on Jan Hulstijn’s essay examines the importance of processing pressures in usage-oriented approaches to language and learning. I use the syntax of negation to illustrate this line of research and its relevance to linguistic inquiry in general.

1. Introduction

In his lead article on Basic Language Cognition (BLC), Jan Hulstijn emphasizes the need for simplicity in linguistic theory and for research on a broad range of phenomena—both commendable goals. To them, I would like to add a third factor worthy of attention: the role of processing pressures in shaping language and learning. I will focus my remarks on negative concord, a phenomenon whose unusual properties are highly instructive.

2. Negative Concord

Negative concord is a structural pattern characterized by the use of two separate negators to express simple negation. As illustrated in the examples below from Spanish, Polish and Korean, one of these negators is typically a sentential negative (equivalent to English not) and the other is a negative pronoun (e.g., similar to nobody, nothing, etc.).

Spanish
Juan	no	hizo	nada.
Juan	not	did	nothing
‘Juan did nothing.’

Polish (Zeijlstra, 2004, p. 124)
Janek	nie	pomaga	nikomu.
Janek	not	help	nobody
‘Janek helps nobody.’

Korean
(Nan)	amwuto	an	po-ass-ta.
I.topic	nobody	not	see-pst-decl
‘I saw nobody.’

Negative concord was also a staple of Middle English, and has survived in a number of non-standard varieties of Modern English, both in the United Kingdom and in the United States (see, e.g., Robinson & Thoms, 2021).1

Middle English
Ne	taketh	nothing.
not	take	nothing
(= ‘Take nothing.’/‘Don’t take anything.’)

Non-standard Modern English

I didn’t do nothing.

(= ‘I did nothing.’/‘I didn’t do anything.’)

The existence of negative concord in human language raises two puzzles—one typological and the other developmental.

The typological puzzle involves the pattern’s apparent redundancy, stemming from the use of two negatives when one could evidently suffice (cf. standard English I did nothing).

If the meaning of [negative-concord sentences] is equivalent to that of the structure with a single negation …, why do we need them? Negative concord appears to be seriously redundant.

(Giannakidou & Zeijlstra, 2017, p. 10)

Intriguingly, negative concord is nonetheless very common. Indeed, it appears in 170 of the 206 languages surveyed in the World Atlas of Linguistic Structure (Haspelmath, 2013), raising a question for all theories of language. Why would a phenomenon as fundamental as negation not be expressed in a simpler and apparently more efficient way?

The facts of linguistic development reveal a further puzzle. Some children learning English as a first language go through a stage in which they employ negative concord even though it is not encountered in the speech of those around them. The tendency was first reported by Bellugi (1967), who found that one of the three English-speaking children who she studied (Adam) regularly produced negative-concord sentences even though he had not been exposed to such patterns in the speech of his parents (Brown, 1973, p. 381; Thornton et al., 2016, pp. 7–8).

I didn’t do nothing.

(file 63, age 3;5)

I didn’t call him nothing.

(file 72, age 3;8)

Because nobody didn’t broke it.

(file 107; age 4;5)

A much larger study, conducted by Hein et al. (2022), examined speech samples from seven children and their caregivers. Among the 328,972 utterances produced by the children, 909 were sentences that contained negative items such as nobody, nothing, never or no. Of these, 178 (19.6%) also contained a sentential negative, as in the examples below.

We don’t want no gas.

(Adam 3;11)

No tigers don’t bit you?

(Mark 2;08)

I don’t care about nothing.

(Ross 5;04)

No one’s not drying him, mum.

(Fraser 3;00)

Crucially, this innovation cannot be attributed to usage per se, as there were no examples of negative concord in the speech of the children’s caregivers.

How then are we to make sense of the typological and developmental facts that underlie negative concord? A promising line of inquiry involves the role of processing in shaping the properties of language and the manner in which they are acquired.

3. Biuniqueness

Elsewhere, (O’Grady, in press) have suggested that the properties of negative concord are best understood by reference to a processing factor known as ‘biuniqueness’ (e.g., Dressler, 2000, p. 291).2

Biuniqueness

A particular form has a particular function and vice versa.

As a preliminary example, let us consider the expression of plurality in English, in which the noun suffix -s almost always marks the plural and the plural is almost always marked by -s.

-s	<===>	plural
dog		dogs
book		books
tree		trees
apple		apples
, etc.

Such one-to-one correlations facilitate processing by allowing a quick and reliable mapping between the two. Tellingly, where exceptions exist (sheep, teeth), there is a chance of over-regularization by language learners.

two sheeps

my tooths

‘Errors’ of this type confirm the preference for the one-to-one relationship between form and function that defines biuniqueness.

Turning now to negation, the following three options are relevant.

not/n’t	<===>	basic negation	The children didn’t leave.
nobody	<===>	null set of people	Nobody sang.
nothing	<===>	null set of things	They ate nothing.

As the examples illustrate, all three options can negate a sentence. If the children did not leave, there was no leaving event. If nobody sang, there was no singing event. And if nothing was eaten, there was no eating event.

Of the three negative patterns above, the first turns out to be the most commonly used by far. In the BabyLM corpus assembled by Warstadt et al. (2023), for example, there were 126,611 instances of not or n’t, which occurred in approximately one of every ten sentences.3 In contrast, there were a mere 962 instances of nothing, nobody or no one—less than one percent of all negative sentences.

These facts underlie a developmental path that can explain both children’s early attraction to negative concord and, relatedly, its widespread presence in the world’s languages. The first step on that path is triggered by frequent exposure to the use of not to express sentential negation, as already noted. In accord with biuniqueness, a strong relationship between form and function emerges.

not

<===>

sentential negation

A second step involves the reaction to those (relatively few) sentences in which a negative pronoun is responsible for negation (e.g., They did nothing). Given the very strong association between not and sentential negation, the co-occurrence of that negator with a negative pronoun is hardly surprising (any more than the use of -s to pluralize sheep would be). Indeed, that is exactly what produces negative concord.

Nobody	don’t	like that.
↑	↑
pronoun	sentential negative is
signals	triggered because the
negation	sentence has been negated

A third and final step is corrective. As learners of English come to observe that a negative pronoun can occur without an accompanying not, they abandon the negative-concord strategy.

Nobody don’t like that.

>

Nobody don’t like that

>

Nobody likes that.

This scenario allows for variation among children with regard to when and in what quantity they encounter negative pronouns in ambient speech. Early and frequent exposure to sentences containing a negative pronoun but no sentential negative (Nobody sang, I did nothing) could well limit the duration of the negative-concord stage or even circumvent its occurrence entirely. On the other hand, infrequent exposure to such patterns could encourage an extended period of negative-concord use, as reported by Hein et al. (2022).

4. Concluding Remarks

Negative concord manifests two puzzling properties: typological frequency despite its seeming redundancy and, even more surprisingly, its early use by English-speaking children despite its absence in the speech of their caregivers. Both properties can be traced to the importance of processing factors in converting usage data into an actual syntactic system. In the case considered in this brief essay, usage serves at least three functions:

• Usage provides data that can lead to the creation of particular form–function relationships, including the one that underlies negative concord.
• Usage can confirm the adoption of negative concord in languages such as Spanish, Polish and Korean, where this pattern is in fact called for.
• Usage can help suppress the use of negative concord (in cases like English) by presenting children with sentences in which a negative pronoun occurs without not.

Crucially though, the ultimate source of these effects can be traced to biuniqueness, a factor designed to reduce processing cost. This is ultimately why negative concord is a viable strategy for expressing negation, why it is so common in the world’s languages, and why it can be manifested in the speech of children who never encounter its use.

As things currently stand, there is widespread agreement—perhaps even a consensus—that cognition, usage and processing each have a decisive role to play in understanding the workings of language. Indeed, it is not unreasonable to think that the future of linguistics may well be forged in debates over the precise contributions of those factors to the evolution and acquisition of language. BLC is well positioned to explore these issues as part of its agenda, based as it is on a commitment to understanding both typological variation and developmental trajectories. I look forward to following the findings of that research.