Phonological Interference Research Articles

Reading comprehension in the presence of unattended speech and music

This series of experiments investigated whether the detrimental effects of unattended speech that have been obtained in short-term memory tasks would be obtained in reading comprehension. Such effects would be expected if reading comprehension depends on the maintenance of phonological information in short-term memory. The first three experiments demonstrated that unattended speech but not music interfered with reading comprehension while unattended music had a greater interfering effect than speech on a music identification task. Experiments 4 and 5 showed that the detrimental effect of the speech backgrounds on reading was due to their semantic rather than their phonological properties. The failure to find a phonological interference effect argues against a role for phonological short-term memory in reading comprehension.

When words collide: Orthographic and phonological interference during word processing

The interaction between orthographic and phonological information was studied in two experiments by requiring subjects to match visually presented word pairs on the basis of their visual or rhyming similarity. Word pairs either rhymed and looked alike, rhymed but did not look alike, looked alike but did not rhyme, or did not rhyme and did not look alike. In Experiment 1 under rhyme matching, reaction time (RT) was markedly increased whenever there was a conflict between orthographic and phonological cues. Under visual matching, overall RT was shorter than rhyme matching, with visually similar rhyming and non-rhyming pairs producing equally rapid and short responses compared to the non-rhyming but visually different word pairs. Most subjects also responded slower to rhyming and visually different stimuli compared to word pairs that did no look alike or rhyme. Experiment 2 sought to specify the processing locus of these effects by recording event-related brain potentials (ERPs) under task conditions similar to the first experiment. The RT data essentially replicated the effects found in Experiment 1 for both matching tasks. The ERP data viewed in the context of these results suggested that the interaction of the orthographic and phonological codes begins at least at the stimulus comparison processing stage, but that the conflict may also contribute to delays in response selection. The results are discussed in terms of several current models of word processing.

RELATIONSHIPS BETWEEN SPEECH AND READING AMONG SECOND LANGUAGE LEARNERS

This paper describes the results of a short‐term memory, reaction time study with students of English as a second language. Thirty‐five foreign students at the University of New Mexico did a series of short‐term memory tasks designed to determine if phonological encoding of visually presented verbal material occurred during reading. The memory tasks were done in two conditions: simple visual presentation of material (nonshadowing condition) and visual presentation of material accompanied by an oral repetition task (shadowing condition). A similar study done by Kleiman (1975) with native speakers of English showed that reaction times increased (slowed down) in the shadowing condition. In the present study the second language students' reaction times decreased (speeded up) in the shadowing condition. The explanation offered for this result is that the second language students in the nonshadowing condition show phonological interference from their native language which slows reaction times; however, the shadowing task which involves the oral production of English curtails the interference, thereby improving the reaction times. In addition to the short‐term memory tasks, the students in the study took a reading proficiency test (the Nelson‐Denny Reading Test). Correlations between the results of the memory task reaction times and reading proficiency scores showed few significant results.

Phonological and tactual coding of Braille by blind children.

Two experiments were carried out to investigate phonological and tactual coding in Braille reading by blind children. In the first, the children read aloud two lists of word pairs, one item at a time. The second or 'target' word of each pair (e.g. on) was the same in both lists. In one list, the congruent list, the phonology and orthography of the first word of the pair (sun) provided no inconsistency with respect to the target word (sun-on). In the other incongruent list, the preceding word was similar in orthography but inconsistent in phonology with respect to the target word (son-on). The children named the target words (i.e. on) significantly faster in the context of the congruous list than in the context of the incongruous list, thus revealing a phonological effect in the blind children's reading of single words. In addition, direct lexical access, from tactual input, seems to proceed with the same facility for the blind as does visual input for the sighted. In the second study the children read aloud words which were either orthographically regular or irregular. The results indicated that the irregular words took longer to name than the regular words. It was suggested that blind children, like sighted children, show phonological interference effects in word naming. The results are discussed in relation to recent models of the processes involved in word naming.

Does linguistic reversibility exist? an experimental exploration of a psycholinguistic construct

Three experiments were devised involving 176 subjects aged between 5-0 and 8-1 years to examine children's receptive control of a set of four intersentence relations (identity, antonymy, structure reversal, equivalence) and to relate this linguistic ability to levels of cognitive functioning as assessed by the Piagetian measureDeux Sens de l'Orientation. The results indicate a significant but moderate correlation between performance on the two tasks. Cognitive precedence with respect to corresponding linguistic ability was observed in only one of the experiments, where linguistic performance was unsually low due to phonological interference between the lexical items used. The results were interpreted as giving evidence of linguistic reversibility—partially independent from cognitive reversibility though isomorphic with it—in the perceptual-figurative mode, resulting in flexible and efficient processing of linguistic inversions and even converse relations by preoperational children. These data complement previous psycholinguistic research in the Piagetian tradition, which had shown cognitive precedence relative to corresponding linguistic processes. It is suggested that previous research had tapped mainly linguistic processes in the cognitive operational mode, and that future research needs to examine the interrelation between the two modes.

Interference and the persistence of articulatory responses.

An experimental test of the effect of phonological interference on short-term recall was conducted using non-English phonological sequences which were easily pronounceable. Subjects were normal speaking and normal hearing third-grade English speaking children. Experimental subjects produced the phonologically inadmissible [3a], [u'mI], [vepsilon], and control subjects produced the phonologically allowable [d3a], [u'mî], [veI]. Following a retention interval of 30 sec, one free recall trial was administered. Experimental subjects correctly recalled significantly less often than control subjects, suggesting that phonological rules can effect a decrement in short-term recall for phonetic units which are easy to pronounce.

A human sound transducer/reproducer: Temporal capabilities of a profoundly echolalic child

Stimulus-echo latencies of a blind, severely echolalic child were measured to assess the capabilities of spontaneous transduction/reproduction of a verbal signal. Mean offset-to-onset latencies of 270 msec and onset-to-onset latencies of 792 msec are comparable to laboratory reaction times for syllable imitation. Results are considered within the theoretical context of physiological speech, i.e., externalized exposure of a psittacine level of verbal processing divorced from semantic, syntactic, lexical and, quite probably, phonological interference.

Do Adjectives and Functors Interfere as Much as Nouns in Naming Pictures?

EHRI, LINNEA C. Do Adjectives and Functors Interfere as Much as Nouns in Naming Pictures? CHILD DEVELOPMENT, 1977, 48, 697-701. Thirdand sixth-grade readers were asked to label sets of pictures printed with distracting words (either nouns, adjectives, or functors) and nonsense syllables. Picture-naming latencies were significantly longer with nouns than with adjectives and functors, indicating that noun interference stems from semantic-conceptual rather than phonological sources. Latency patterns among the words were similar for younger and older readers, suggesting that both groups are equally susceptible to semantic sources of interference. Adjectives but not functors yielded substantially longer latencies than nonsense syllables. However, interpretation is difficult because of the inadequacy of nonsense syllables as a control for phonological interference.

Dialectal Differences and Learning Disorders

Sufficient information has accumulated to indicate that there is frequently a positive relationship between functional mental retardation and linguistic conflicts where speakers use standard English and auditors use nonstandard English. In order to negate the effects of this relationship, phonological, grammatical, and lexical interferences must be reduced between subculturally different teachers or clinicians and children. Ear training methodologies are suggested that provide for auditory contrastive analysis or bidialectal teaching. The inappropriateness of conventional auditory discrimination testing and training programs is discussed.

An Investigation of Phonological Interference

1. The ease or difficulty of learning phonological categories, experienced by a speaker of language X attempting to learn language Y, has been attributed by linguists to (1) the competing phonemic categories of native language (N) and target language (T) systems, (2) the allophonic membership of these phonemic categories, and (3) the distributions of these categories within their respective systems. It has been noted that the higher the degree of similarity between the N and T phonological categories, the easier it is for the speaker to learn the T phonological categories, and the converse has been held to be true. In line with Hockett's definition of the phonological system of a language (1958:24) as 'not so much a set of sounds as it is a network of differences between sounds', it has been assumed by linguists that it is this very existence of a SYSTEM of distinctive and nondistinctive features which causes interference when the speaker of one language attempts to learn another language in which the phonological system is composed of partially similar and completely different distinctive and nondistinctive features. It has further been believed that the speakers of a language learn to attend only to those features which are distinctive and to ignore those which are redundant. Bloomfield says (1933:79), in reference to distinctive features, that 'the speaker has been trained to respond only to these features and to ignore the rest of the gross acoustic mass that reaches his ears.' Once a speaker has learned to attend to certain features only, he supposedly approaches all other languages through his own 'grid' of distinctive versus nondistinctive features (Troubetskoy 1955:54). Since the speaker of N will interpret the phonological system of T in terms of the features of N, interference is expected. Wolff (1950:38), in discussing a hierarchy of interference encountered by speakers of Puerto Rican Spanish learning English, notes that it is easier for everyone to learn a completely new phoneme which does not exist in his native language-e.g. for a speaker of Puerto Rican Spanish to learn American English (AE) /r/ or /j/-than it is to learn a partially similar class in the target language that will involve negative transfer caused by the N system-e.g. AE /p/, with aspirated allophones, as opposed to Puerto Rican /p/, which exists only as a nonaspirated stop. Weinreich (1953), through a contrastive analysis of Romansch and