tl;dr-ELT

You might picture bilinguals as having two separate languages stored in their minds. Vocabulary certainly differs across languages, but what about the underlying grammatical machinery? When an English-Spanish bilingual says cat/cats or gato/gatos, does the brain switch between two different systems, or use the same one?

A fascinating recent study by Xuanyi Jessica Chen & Esti Blanco-Elorrieta (New York University), published in The Journal of Neuroscience, tackled exactly this question by examining how bilingual brains handle one of the most basic grammatical operations: making words singular or plural.

The Study

The researchers recruited 23 highly proficient Spanish-English bilinguals aged 18-45. Participants completed a phrase-production task while their brain activity was recorded using magnetoencephalography (MEG), which tracks neural activity with millisecond precision.

Participants saw nouns on a screen and heard cues such as one, two or say. Depending on the cue, they either:

• repeated the word exactly as shown
• changed it from singular to plural
• changed it from plural to singular

For example:

• boat + two → boats
• boats + one → boat
• boat + say → boat

The design was particularly clever because it separated several factors that are usually tangled together:

• grammatical processing
• meaning (one vs many)
• sound changes
• language (English vs Spanish)

The researchers also included:

• cognates (e.g. similar words across languages)
• non-cognates
• pseudowords (invented words that follow English & Spanish sound patterns)

This allowed them to test whether grammar relies on stored vocabulary or on more abstract mental processes.

The Findings

Perhaps the most striking finding was that the same left frontotemporal brain network was recruited for grammatical inflection in both English & Spanish. Brain activity patterns generalised successfully from one language to the other, suggesting that bilinguals are not using separate grammatical systems.

Even more remarkably:

• The shared neural pattern appeared extremely quickly, beginning around 100 ms after the grammatical cue.
• English & Spanish showed very high overlap in the brain regions involved, with overlap indices of 0.94 (frontal regions) & 0.91 (temporal regions).
• The same neural mechanism worked for different plural types, including words whose singular & plural forms are identical.
• The mechanism also worked for pseudowords, suggesting that grammatical processing does not depend on stored lexical entries.
• Cognates offered no special advantage. Sharing form across languages did not strengthen the grammatical effect.

One way to think about this is that the brain seems less concerned with specific words than with the abstract operation being performed.

Imagine learning a newly invented noun such as flim. Most speakers can instantly create flims despite never having encountered the word before. This study suggests that the brain’s grammar system is designed precisely for this sort of productive rule application.

Another interesting finding was that Spanish triggered the shared grammar network roughly 110 ms earlier than English. The authors suggest this may reflect Spanish’s highly regular plural system rather than any fundamentally different grammatical machinery.

Why This Matters

These findings sit within a broader shift in bilingualism research.

Older views often assumed that bilinguals maintained largely separate language systems. More recent accounts, such as the Integrated Language System model proposed by Blanco-Elorrieta & Caramazza, argue that bilingual language knowledge is deeply interconnected.

The current study provides some of the strongest time-sensitive neural evidence yet for that position. Rather than storing separate grammatical engines for each language, bilingual brains appear to reuse the same underlying computational machinery whenever equivalent grammatical operations are required.

The findings also align with research on:

• cross-linguistic syntactic priming (Hartsuiker et al., 2004)
• shared bilingual syntax (Declerck et al., 2020)
• universal language networks across many language families (Malik-Moraleda et al., 2022)
• productive, rule-based accounts of morphology associated with Steven Pinker & others

Teacher Takeaways?

• Learners may not be building entirely separate grammatical systems for each language. Existing grammatical resources could be reused across languages more than we sometimes assume.
• Grammar seems to be highly productive rather than tied to memorised word forms. This may help explain why learners can often apply rules to unfamiliar vocabulary.
• Persistent learner language patterns may reflect how existing grammatical resources are being adapted to new linguistic contexts rather than simply a lack of knowledge.

If the brain can apply the same grammatical operation across languages, where do you think the boundary between “one language” & “another language” really lies?