Can brain waves show a language changing before speakers do?
For verbs whose grammar is shifting in Icelandic, readers' brain responses already treated the new word order as normal, even though their conscious judgments still partly favoured the old rule.
Source
Language Processing as a Precursor to Language Change: Evidence From Icelandic
Study at a glance
- Design
- Human experiment — Within-subject EEG experiment: participants read Icelandic relative clauses with three verb types (active, alternating, dative-subject experiencer) whose post-verbal noun was nominative or dative, then judged acceptability and answered a comprehension question
- N
- N=23 · 23 right-handed native Icelandic-speaking university students; seven earlier participants run with a different, acceptability-only protocol were excluded
- Population
- Students at the University of Iceland, Reykjavik
- Outcome
- Single-trial ERP amplitudes in the N400 and late-positivity windows at the critical noun; acceptability judgments and comprehension accuracy
Structured fields used in claim comparison tables when every cited study has a complete layer.
What they did
Icelandic, like Old English, marks who does what to whom with case endings, and some of its verbs seem to be drifting toward the English-style pattern where the subject is nominative and comes first. Native Icelandic students read sentences word by word while their EEG was recorded; the sentences used active verbs (already fitting the new pattern), alternating verbs (allowing both orders) or dative-subject experiencer verbs (requiring the old, dative-first pattern). The key noun was marked either nominative or dative, so the sentence followed a nominative-before-dative or a dative-before-nominative order. After each sentence participants judged whether it was acceptable and answered a question about who did what.
What they found
For active verbs, the dative-first order produced a clear N400 followed by a late positivity, the classic brain signature of an unexpected structure, and was consistently judged unacceptable. Alternating verbs showed a similar but weaker pattern that emerged over the course of the experiment, with more variable judgments. For experiencer verbs, which prescriptively require the dative-first order, judgments were highly variable and brain responses did not distinguish the two orders at all, suggesting case marking had become uninformative during online reading. Late-positivity amplitude predicted trial-by-trial acceptability better than N400 amplitude, and overall comprehension accuracy was 75%.
The limits
What it doesn't show
The study is a single snapshot with a small student sample and no longitudinal or historical data, so it cannot show that these brain responses actually precede or predict future change; the authors present that as a hypothesis. No production data were collected, so it is unknown whether participants who accept or process the new order also use it themselves. Trial numbers differed across verb classes and the effects for alternating verbs changed during the session, possibly influenced by the many active-verb sentences. Interpreting a null ERP difference as 'case has become uninformative' rests on theoretical assumptions about what the N400 reflects.
Key terms
- Event-related potential (ERP)
- The average EEG voltage change time-locked to a specific event, such as reading a word, used to track processing millisecond by millisecond.
- N400
- A negative ERP deflection around 300 to 500 milliseconds after a word, larger when the word is unexpected; here interpreted as a prediction error.
- Late positivity (P600)
- A later positive ERP deflection often seen after grammatical or structural anomalies, linked here to evaluating the sentence.
- Case marking
- Changing the form of a noun (e.g., nominative vs. dative) to signal its role in the sentence, instead of relying on word order.
- Predictive coding
- The idea that the brain constantly predicts upcoming input and registers prediction errors, weighted by how precise the prediction was.
Flashcards
0 of 10 answers reviewed
Research intelligence for this paper
See its role on concept claims, tensions it is part of, placement history, and related discoveries.
Quiz yourself
Which verb type prescriptively requires a dative-before-nominative order in Icelandic?
Common questions
Why study Icelandic to learn about language change?
Icelandic combines rich case endings like Old English with fairly fixed word order like modern English, and some of its case patterns appear to be shifting, so it offers a living example of a transition English went through centuries ago.
Why do the authors think the N400 may be a better early warning sign than the late positivity?
The late positivity tracked people's conscious acceptability judgments closely, whereas the N400 is argued to reflect automatic predictions during reading that may shift before conscious judgments or prescriptive grammar do.
If people understood active-verb sentences correctly in both orders, why was one judged unacceptable?
Comprehension depended mainly on which noun sat in the subject position, much as in English, so case endings mattered little for understanding even when they made the sentence sound wrong.
More on Language