Do Beez Buzz? Rule–Based and Frequency–Based Knowledge in Learning to Spell Plural –s

Abstract

There has been much discussion about whether certain aspects of human learning depend on the abstraction of rules or on the acquisition of frequency-based knowledge. It has usually been agreed, however, that the spelling of morphological patterns in English (e.g., past tense -ed) and other languages is based on the acquisition of morphological rules, and that these rules take a long time to learn. The regular plural -s ending seems to be an exception: Even young children can spell this correctly, even when it is pronounced /z/ (as in bees). Reported here are 3 studies that show that 5- to 9-year-old children and adults do not usually base their spellings of plural real-word and pseudo-word endings on the morphological rule that all regular plurals are spelled with -s. Instead, participants appeared to use their knowledge of complex but untaught spelling patterns, which is based on the frequency with which certain letters co-occur in written English.

Children's Implicit Learning of Graphotactic and Morphological Regularities

“…Furthermore, there was no trend toward a reduction in the magnitude of the graphotactic regularities impact in the GM condition from the second to the fifth grades in Study 1, and the impact of the graphotactic regularities was still observed among adult participants in Study 3. This echoes Kemp and Bryant's (2003) study in which children and adults correctly used -s for plural pseudo-words more often when the final /z/ was preceded by a consonant than when it was preceded by a long vowel, despite the possibility of relying on a rule specifying that plural nouns are spelled with -s.…”

“…For example, when words such as éléphanteau (baby elephant) are encountered in a context in which it is associated with the feature of diminutiveness, an association could be established between the diminutiveness and /to/ or teau rather than between the diminutiveness and /o/ or eau. According to this account, participants would use eau when they are told that a little /klat/ is a /klato/, but they would be less inclined to use eau when they are told that a /klafo/ is a little /klaf/ because /fo/ does not sound like a diminutive (see Kemp & Bryant, 2003, for a similar interpretation). Note also one could argue that participants learn rules that are specific and idiosyncratic rather than general (e.g., /to/ is spelled teau when it is a diminutive suffix).…”

Grammatical understanding, literacy and text messaging in school children and undergraduate students: A concurrent analysis

“…For example, some word suffixes (inflections) are always spelled in the same way despite differences in pronunciation, to reflect their shared grammatical structure (e.g., the -ed inflection of mixed, stirred, and kneaded). Children often start spelling these patterns just as they sound (stird for stirred, keez for keys) and have to learn grammar-based spelling consistencies (Kemp & Bryant, 2003;Nunes et al, 1997 Kemp, 2012;Rosen et al, 2010;Varnhagen et al, 2009). It is also common to omit punctuation, to use multiple exclamation marks or question marks, or to use symbols such as emoticons, often in place of conventional punctuation (De Jonge & Kemp, 2012;Provine, Spencer, & Mandell, 2007;Rosen et al, 2010).…”

Statistical approaches to language acquisition and the self-organizing consciousness: a reversal of perspective

“…For instance, transfer to a new situation in which a rule applies is never as perfect as a rule-based theory would predict (e.g., Pacton, Perruchet, Fayol, & Cleeremans, 2001), because behavior is sensitive to the distributional properties of the original input. Even in conditions where a simple rule exists, and is taught at school, such as the ''s'' to mark regular plurals in English, participants do not seem to rely on it (Kemp & Bryant, 2003). To account for these observations, rule-based theories need to add to their (costly) postulate for rule guidance other (costly) postulates for mechanisms preventing the application of rules 2 (for an illustration about Marcus' account of past-tense formation, see Bybee, 1995, p. 448) Thus, recent research demonstrates that a distributional approach to language is now gaining increasing plausibility.…”