Speech Error Classification: An exploration of word class, semantic links, and exceptions to the rules

Speech errors pervade common language usage in almost every context. Such a present aspect of language usage has been found by a number of linguists to provide interesting insight in the mechanics of speech production. The works of Nooteboom (1969), Fry (1969), and those preceding them have provided a wide ranging number of studies and literature about the means of how language is formed in the mind, and articulated – something that has been provided by observing the broad variety of ways that production might fall short in some way or another. However, it is of great import that one should first establish some of the defining features of specific speech errors, so as to investigate the intricacies of the field of Speech Error Analysis. To this end, the research and analysis found in this paper will be a matter of testing the categorical boundaries that are drawn around three particularly common classifications of error – those which Nooteboom (1969) refers to as Selection Errors: The Lexical Selection Error; The Blend; The Malapropism. Nooteboom’s definitions of each of these will be compared to a corpus of 154 naturally occurring selection errors, extracted from a greater corpus of 272 general speech errors, a self gathered list which includes syntactic and phonological shortcomings in speaker’s utterances as well. An analysis of the data collected has indicated that the most common of all kinds of speech error is the Lexical Selection Error (Henceforth LSE) – an often semantically linked yet erroneously selected word in the place of what is initially meant. This being perhaps the most archetypal classification of Nooteboom’s selection errors, the wealth of LSEs began to indicate an adherence to a word-class rule as suggested by Nooteboom (1969). This rule dictates that nouns, for example, are replaced by erroneous nouns, and mutatis mutandis, the same is true across all word classes and, as a consequence, speech errors appear to be bound to their targets in such a way. Nooteboom’s application of this rule was applied not only to LSEs but also to Selection Errors as a whole, a term applicable to Blends as well as Malapropisms and so one should expect similar results. While the data at hand harbours far fewer blends and malapropisms than LSEs there seems to be less of an adherence to such a rule in these two classifications. Fromkin (1971) verifies the research of Nooteboom in her work with her own data, in such a way that the current research seeks to replicate. The indication from initial analysis indicates some particular blends, malapropisms, and LSEs that don’t seem to abide by Nooteboom’s word class rule. This will be further explored in such a way as to see if these exceptions are grounds to question these means of classification. Fry, D.B (1969) The Linguistic Evidence of Speech Errors. Speech Errors As Linguistic Evidence (1973), p.157. Fromkin, V. (1971) The Non-anomalous Nature of Anomalous Utterances. Speech Errors As Linguistic Evidence (1973) p.215. Nooteboom (1969) The Tongue Slips into Patterns. Speech Errors As Linguistic Evidence (1973) p.144.