The relevant verb, doxazein, often means 'to judge that something is so-and-so.' Hellenistic discussions envisage three attitudes that cognizers take to impressions (how things seem to them): assent, rejection, and suspension of judgment (epochê).
The entrenchment measure is simply the (natural log N + 1) frequency of the bare (i.e., without un-) form of the relevant verb in the BNC (all texts).
This study tests the prediction of the entrenchment hypothesis of a negative correlation between the rated acceptability of un- prefixation errors (e.g., * unsqueeze, *unfill) and the overall frequency of the relevant verb (e.g., squeeze, fill, in all forms except with the prefix un-) in a representative corpus.
This Court rejects the most natural grammatical reading, adopted by the Ninth Circuit, under which "knowingly" modifies only the relevant verbs in subsections (1) and (2), and does not extend to the elements of the minority of the performers, or the sexually explicit nature of the material, because they are set forth in independent clauses separated by interruptive punctuation.
Wattarujeekrit et al. (2004) [27] apply a lexical sampling method (Palmer et al., op cit, [2] p. 85) to construct predicate-argument representations for a small number of biologically relevant verbs.
As we used a positive list of relevant verbs, this reduction regarding the number of generated PASs would help to reduce the amount of data to be managed but would not improve the accuracy of the proposed approach.
In contrast, many of the relevant verbs, and the un- prefixation construction itself, are relatively rare.
Recall that relevant verbs (e.g., appear) boost the activation of all constructions in which they appear even those that score low on relevance (e.g., [SUBJECT] [VERB])—in proportion to their frequency of co-occurrence.
Even for simpler formulations based solely on verb frequency, the relative scarcity of many of the relevant verbs means that entrenchment will presumably remain unreliable until a very large volume of data has been encountered.
The semantic verb class hypothesis works well when the relevant verbs cluster tightly into semantically based classes such as verbs of "contained motion taking place in a particular manner (e.g., slide, skid, float, roll, bounce)" (Pinker, 1989: 1303), which may appear in both the intransitive and transitive construction (e.g., The ball rolled; John rolled the ball).
Item-in-construction frequency: The activation boost received by highly relevant verbs (e.g., appear, materialize, vanish) spreads to the constructions in which those verbs have been previously encountered, in proportion to the frequency with which the verb has appeared in each.
