Phonetic Representations in the Mental Lexicon (Note: all quotations are taken from Chapter 4 by John Coleman) Carlos Sarmiento Caroline Avilés ENGL 6111

What is the Lexicon? The lexicon is the knowledge that a native speaker has about a language. This includes: The form and meanings of words and phrases. The appropriate usage of words and phrases. Relationships between words and phrases.

Arguments Coleman argues that there is no material evidence for abstract phonological representations and units of the kind. The representations of the word forms are phonetic not symbolic-phonological. The mental lexicon has many kinds of phonetic representations including: auditory, articulatory, and visual.

Arguments If the notion of non-abstract, superficial phonetic representations is to be believed, there must be some phonological phenomena which isn’t taken into account by using phonetic representations. Coleman argues that certain facts regarding sociolinguistic variation, gradual phonetic change and child language acquisition are problematic for parsimony of representation.

Syllables An observation was made that some words mismatch in the number of syllables and the number of sonority peaks. Some words consisting of a single syllable may have more than one sonority peak. Coleman argues that any or all of [mIlƏk], [noƱƏn] and [fɑɿƏl] are phonetically disyllabic realizations of the phonological (i.e. lexical) monosyllables /mIlk/, /fɑIl/, / noƱn/.

Syllables Speakers who hear these words have to make an interpretation. If this was an entirely bottom-up process, they might interpret [fIlƏm] as ‘fill’em, which would give the speaker problems of syntactic and semantic processing. Due to the previous knowledge by the speakers it leads them to interpret the word as [film].

Syllables Evidence shows that nine month old children are more attuned to the phoneme transition probabilities of their ambient language. Analyzed data show that eight month old babies segment a continuous stream of spoken syllables into word-like units. This evidence allows a more data orientated approach to language acquisition rather than the more speculative “top-down” proposals.

Previous Knowledge New words may be generated by words already stored in the lexicon. These words may be partly understood by referencing known words. Even made up words are likely to be accepted by the speaker-hearer as long as they follow a set pattern already stored in the lexicon. The lexicon will automatically suggest word stressing using the mental storage of words.

Phonemes An argument presented by Coleman ponders why if phonemes are real, many students have difficulty determining how many phonemes are in orthographic “ng” “ch”. “Are diphthongs and affricates one phoneme or two?” A study by Sapir shows that there are speakers who insist that there is a presence of sound in words where none was supposed to be heard thus giving evidence that there is a latent phoneme in the speaker’s lexical representation.

Findings Statistical studies confirm that literacy has an impact on spoken language. Coleman argues that “auditory representations are central to word production as well as recognition, whereas articulatory representations are employed only in speech production.”

Stress The stress system has shown to have remarkable stability. When stress changes (in a language) over time, there is evidence that the speaker continues to use the older pattern of stress placement. What this suggests is that phonological representations are phonetic memories.

Speech Perception and Production Studies Evidence shows that speaker-hearers are aware of finer phonetic differences than phonological theory accounts for. Under appropriate circumstances hearers may employ features to resolve a contrast between similar sounds such as: /d/ and /t/. Example: kitty-kiddy It could be suggested that speaker specific details of spoken words are stored together with the usual phonological information.

Speech Perception and Production Studies Subjects identify more accurately when they are spoken by a familiar speaker than an unfamiliar one which shows that people have acute auditory memories of words stored together with the linguistic information. Coleman concludes that “very specific pronunciation details, such as speaker-specific detail and word frequency, are encoded in the lexicon.”

Effect of Word Frequency A number of works by Hooper and Bybee shows that word sounds may change due to frequency. This can be clearly seen in the Spanish language with the use of /s/ which is regularly omitted. Ex. Las otras, las únicas.

Conclusions The cognitive model of phonology needs associations between auditory and articulatory representations, semantic and articulatory, auditory and visual/orthographic, etc. Instead of searching for a single kind of word forming representation, evidence shows many different types of modality-specific representations, and various paths of association between them.