Bibliographic record
Abstract
Imagine listening to a language you don't know. When does one word end, and another begin? Human infants face a similar challenge, yet remarkably, they grasp the structure of their mother tongue naturally without receiving any explicit indications. By six months, they recognize some common nouns, and by one year, they start saying their first words. This learning begins from birth, with newborns already sensitive to speech patterns. Previous studies have shown that the likelihood of certain syllables appearing after others allows infants to detect regularity and separate speech into chunks. This is because some syllables are more predictive of what comes next than others. For example, in English, many different syllables can follow ‘the’. However, it is highly likely that ‘brocco’ will be followed by ‘li’. The ability to detect these regularities is known as statistical learning. However, whether this relies on a general mechanism or is restricted to a specific speech component, such as the sequence of syllables, remained unknown. To investigate, Fló et al. measured brain electrical activity of newborns up to 4 days old in response to speech specifically designed to contain certain patterns of syllables or voices. In one experiment, the speech had regular patterns in the syllables, while in a second experiment, the pattern was in the voices, and each voice could utter each syllable. Unlike tracking syllable variation, which can help with learning words, voice changes within a word are unnatural and predicting them is not relevant to real-life speech processing. Therefore, if statistical learning in speech is shaped to promote language acquisition, learning should be restricted to syllable patterns. Instead, if statistical learning is a general mechanism, newborns should also detect the patterns in voice. Analysis revealed that newborns were equally capable of discerning regular patterns in syllables despite voice changes and in voices disregarding the syllable that was pronounced. This suggests that statistical learning is a general learning mechanism that can operate across multiple features. Additionally, pseudo-words (those which resemble a real world but don’t exist in the language) were presented to the newborns after they had been familiarised with speech containing either similar syllable or voice patterns. The researchers observed a specific neural response to the pseudowords only when related to syllable patterns. This neural component suggests that only syllabic structures are considered word candidates and processed by a dedicated neural network from birth. Taken together, the findings of Fló et al. reveal insights into how humans process speech when experience with language is minimal, suggesting that statistical learning may have a broader role in early language acquisition that previously thought.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.017 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.008 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".