Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Never forget: L2 speech is someone else’s L1 speech. That means that real L2 speech is like real L1 speech: often unlike how it’s written. English-speakers say I’ll post my letter to Grandpa, and 99 times out of 100 it comes out with post pronounced pos’, and Grandpa pronounced Grampa. The deletion of the sound /t/ in post my, or the assimilation of one sound to the following one, are “casual-speech processes”. Some such processes, including these two, are very common across languages including Slavic languages, of course. Casual-speech processes are supposed to make life easier for talkers. But ease of articulation is not the whole story, because some of these processes appear in only a few languages, though they involve sound sequences found in many languages. Consider /l/ followed by /r/, as in Kilroy or bellringer. English-speakers don’t say Kirroy or berringer as an easier way of saying those words. But in Hungarian that is exactly what happens – /lr/ becomes / rr/ (e.g., balrol ‘from the left’ becomes barrol). Even in two varieties of the same language, adjustments that happen in one dialect may be unknown in the other. English is a case in point. Phrases like idea of or saw a can be said with an /r/ separating the two vowels at the word boundary. This happens in most forms of British English; in most forms of American English it never happens. (Tip: The Beatles’ A Day in the Life – “I saw a flm today, oh boy” – provides a nice clear example of this phenomenon!). So what happens when L2 listeners are confronted with casual speech processes? Annelie Tuinman’s PhD thesis answered this question (Tuinman, 2011; Tuinman & Cutler, 2011; Tuinman, Mitterer & Cutler, 2011). There is both good and bad news. The good news is that insertions, deletions and reductions in L2 speech are no problem at all – as long as the native language has the same process. In fact L2 listeners are very sensitive to exactly how the process works in the L2 and quickly pick up on any differences with the L1. The case study here was German learners of Dutch. These languages both have the /t/-deletion process, as in English, but there is a slight difference – German speakers don’t usually reduce a /t/ that is a verb ending, but Dutch speakers do (so do English speakers! The verb ending in I passed my exam is just as readily reduced as I post my letters). The German Dutch-learners picked up on this small difference immediately and if anything were even more ready than the native Dutch to expect such a /t/ to disappear. And the bad news? That’s when the L2 process is quite unfamiliar to the L1 ear. This case study involved Dutch listening to their L2, English – the British kind of English, with the intrusive /r/ in contexts such as idea of. Such intrusions never ever happen in Dutch, though in Dutch too there can be word boundaries with vowels on each side (e.g., Papa en Mama – en means and). An interesting property of this process is that it can cause ambiguity. Take a sentence like Canada aided the small African country. A word recognition study showed that when Dutch listeners heard this, spoken by a true Brit, the word RAID sprang to their mind. Native British listeners Editor: Valery Belyanin (Kaluga State University)
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.
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 | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 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 | 0.000 |
| Insufficient payload (model declined to judge) | 0.014 | 0.006 |
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 it