The abstractness hypothesis and morphemic spelling

Abstract

The acquisition of orthography in spelling and reading has seldom been the object of linguistic inquiry due to the common misperception that orthography has no place in grammar. Orthography should be accorded the status of a com ponent of grammar, however, since its units are linguistic signs and the con straints on errors in spelling and reading are similar to those in phonology, syntax, and semantics (Luelsdorff, 1986). Systemic deviation from phoneme-grapheme biuniqueness is the major source of error in the acquisition of a native or foreign alphabetic script. Such deviation is graphemically ambiguous if the relation between grapheme and phoneme is many-to-one and phonetically ambiguous if the relation between grapheme and phoneme is one-to-many. A special case of phonetic ambiguity is morphemic spelling. In this paper we study the acquisition of the uniform <ed> morphemic spell ing of the regular allomorphs of the English past tense morpheme, / Id/, /d/, and /t/, in a large sample of German school children in all the grades of the secondary system of education, the Hauptschule, the Realschule, and the Gymnasium. First, whereas we do find consistent improvement from school system to school system, we reject the view that pupils' performance necessarily improves from grade to grade. Secondly, we find fairly consistently better per formance on / Id/ than /d/ and on /d/ than /t/. This distribution we explain by the Abstractness Hypothesis according to which orthographic representa tions which are less abstract are acquired earlier than those which are more abstract, where abstractness is measured in terms of the number of rules relat ing the orthographic representations to the phonetic ones. Thirdly, the Abstractness Hypothesis also predicts the relative frequencies with which these allomorphs are misspelled and the actual major misspelling types of the regular allomorphs of the past tense morpheme themselves. The Abstractness Hypothesis on the acquisition of morphemic spelling invites testing in other areas of morphemic and partially morphemic spelling, such as derivationally related pairs, the possessive, the regular past participle, the third person singular, and the plural.