The effect of online methods on epistemic inference and scalar implicature
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
How is research on semantics and pragmatics impacted by the growing use of online methodologies, and how does the modality of presentation impact our ability to detect and use a speaker's knowledge state in the service of a linguistic inference? In three experiments, we investigated scalar implicatures both in-person and across three online modalities (text, text + pictures, and video) using a task that required participants to monitor contextual information to infer the mental states of speakers (i.e., whether they were knowledgeable or ignorant with respect to stronger alternative statements). In Experiments 1 and 2 we found no consistent differences across modalities in rates of scalar implicatures, and found that participants rarely computed implicatures when speakers were ignorant (i.e., participants were sensitive to a speaker's knowledge state across all modalities). However, in these first two experiments participants were explicitly reminded to monitor the knowledge state of speakers. In Experiment 3, when these reminders were removed, we again found no effect of modality when speakers were knowledgeable, but found a significant effect when speakers were ignorant. In particular, participants were more likely to erroneously compute implicatures when tested in-person relative to when they were tested online with text only, or with text and pictures. These findings suggestf that online methods may in certain cases offer a useful alternative to in-person testing of pragmatic reasoning, but that care should be taken in selecting methods when they probe subtle mental state reasoning. • Online methods reliably test implicatures when speaker knowledge is emphasized. • Implicature rates differ via modality when speaker knowledge isn't emphasized. • In-person tests lead to over-computed implicatures when the speaker lacks knowledge. • Text methods cause more errors in tracking speaker knowledge than other methods. • Errors in text-only settings don't affect implicature rates despite knowledge issues.
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 | 0.001 | 0.001 |
| 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) | 0.000 | 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 it