The Science of Reading. Группа авторов

Чтение книги онлайн.

СКАЧАТЬ

Nonword Pronunciation

      Many studies have examined children and adults’ ability to read aloud novel letter strings such as nust and mave. This task provides information about readers’ knowledge of spelling‐sound correspondences and their ability to generalize beyond the words they already know. This ability is particularly important for beginning readers, for whom every letter string is initially novel. Generating the pronunciation (or covert phonological code) of an unfamiliar letter string can allow it to be recognized via its spoken form, an important mechanism in learning to read (Share, 1995). Nonword pronunciation may seem artificial, but it taps into the same knowledge and processes that are used in everyday reading.

      The bases of our ability to generalize are an important issue in cognition. For many years, generalization was taken as the primary evidence that linguistic knowledge consists of rules (Pinker, 1994). Generalization – pronouncing novel letter strings – is the principal motivation for the grapheme‐phoneme correspondence rules (GPCs) in dual‐route models. If people lacked this capacity, the lexical route would suffice because words could simply be memorized. Whereas the ability to generate the past tense of a nonword such as wug was taken as evidence for a linguistic rule (Berko, 1958), the ability to pronounce it was taken as evidence for GPCs. The connectionist approach offered a novel account of generalization (Rumelhart & McClelland, 1986), a major theoretical advance. A neural network is trained to perform a task based on exposure to examples. The knowledge encoded by the network can then produce correct output for novel (untrained) items (see Seidenberg & Plaut, 2006, 2014, for reviews).

      The GPCs in the Coltheart et al. (2001) model were handcrafted to produce plausible pronunciations of nonwords and so unsurprisingly they produced accurate pronunciations on the items that were tested. However, the model’s nonword performance deviates from people’s when other phenomena are examined. Here, we briefly summarize issues in three areas: 1) consistency effects for nonwords; 2) relative difficulty of word and nonword naming; and 3) length effects for words and nonwords.

       Nonword consistency effects

      Spelling‐sound consistency also affects nonword pronunciation. Glushko (1979) found that naming was slower for nonwords such as mave, which is inconsistent because of have, an atypically pronounced neighbor, compared to nonwords such as nust, whose neighbors are pronounced alike. The effect of word neighbors on nonword pronunciation presents a particularly strong challenge to the dual‐route approach, which holds that nonwords are pronounced by nonlexical rules, independent of word knowledge.

      Coltheart et al. (2001) discussed Glushko’s study in detail, suggesting that consistency effects for words and nonwords could arise from conflicts between the two routes, as in the case of exception words. Simulating such effects in the DRC model requires changing parameters to increase the activation of words in the lexical network to the point where an exception word such as have could influence both gave and mave. Coltheart et al. (2001) noted that this requirement motivated the “cascaded” property of the model, setting the timing of activity in the two routes to allow such conflicts to occur.

      This discussion is odd because the authors did not report any simulations of word or nonword consistency effects employing the proposed mechanism. Their account of the consistency effect in the Jared study was that it was due to the inclusion of exception words, not conflicts between the routes. No simulations of nonword consistency effects were reported. Zevin and Seidenberg (2006) examined nonword consistency effects reported in three representative studies: Glushko (1979), Andrews and Scarratt (1998), and Treiman, Kessler, and Bick (2003). Whereas the DRC model did not reproduce any of these effects, they were correctly generated by a model based on Harm and Seidenberg (1999).

      In summary, the DRC model presented in Coltheart et al. (2001) does not produce consistency effects for words or nonwords. The authors did not test their proposal that the effects can be obtained by changing lexical activation parameters. In experiments with the model, we have observed that increasing the level of activation in the lexical system creates undesirable side effects, including errors in pronouncing nonwords, especially lexicalizations, and exaggerated regularity effects for higher and lower frequency words.

       Relative difficulty of words and nonwords

      In general, words are read aloud more rapidly than nonwords (Forster & Chambers, 1973; Frederiksen & Kroll, 1976). This “lexicality effect” can be seen as an extension of the standard frequency effect: Pronounceable nonwords are essentially very low frequency words. Coltheart et al. (2001) counted the lexicality effect among the phenomena their model simulated correctly. Again, however, the model’s behavior differed substantially from people’s.

      This point can be illustrated by examining the distributions of naming latencies for words and nonwords from a study by Rastle and Coltheart (1999). Words were named faster than nonwords (the “lexicality” effect”), but the distributions of naming times overlap. As in this case, a statistically significant difference between two means does not indicate that all members of one group differed from all members of the other. (See supplemental materials.)

The DRC model’s performance on Rastle and Coltheart’s stimuli also exhibits a difference between the word and nonword means, consistent with the claim that the model reproduces the lexicality effect. However, the distributions are completely non‐overlapping. The model incorrectly predicts that all words from the study should have been named faster than all nonwords. The subjects read aloud simple nonwords such as rop and scup more rapidly than difficult words such as isle and scythe, but the DRC model produced the opposite result. The anomaly occurs because the lexical route yields phonological output more quickly than the nonlexical route even for odd words.

      Accommodating these results again seems to require changing parameters in the model to slow the lexical route or to speed the nonlexical route, which produces undesirable side effects (lexicalizations of nonwords or regularizations of exception words). Moreover, changing parameters in the model to account for one phenomenon spoils the simulations of other phenomena (Seidenberg & Plaut, 2006).

       Length effects for words versus nonwords

      An additional phenomenon serves to illustrate that DRC’s treatment of nonwords yielded a broad range of anomalous results. Coltheart et al. (2001) simulated a widely cited study by Weekes (1997) that examined the effect of length (3–6 letters) on the pronunciation of high‐and low‐frequency words and nonwords. Nonword reading times showed a linear increase with length. The time taken to read lower frequency words also showed a linear increase with length but the slope was shallower. Higher frequency words showed no effect of length.

      For their simulation, Coltheart et al. (2001) collapsed the data for high and low frequency words, which then showed no reliable effects of length, in contrast to nonwords. The DRC simulation reproduced this word‐nonword difference, because the nonlexical route used for nonwords operates serially, whereas the lexical route does not. The length effect for nonwords but not words was taken as another finding the model successfully explained.

      This account is inaccurate. The DRC model produced length effects for nonwords but not words; however, Weekes (1997) found a length effect for lower frequency words, which was not examined in the simulation. Looking at all three conditions, the correct generalization is about the impact of frequency, not lexicality, on length effects: high‐frequency words < low‐frequency words < nonwords, which are tantamount to very low frequency words. Several studies predating Coltheart et al. (2001) found length effects on word naming that were modulated СКАЧАТЬ