Phonological loop and intermittent activity: A whistle task as articulatory suppression
Abstract The purpose of this study was to examine whether the effect of articulatory suppression is due to the activation of an irrelevant phonology or to intermittent articulatory movements. In the first experiment, subjects were tested for serial recall of visually presented letter sequences that were either phonologically similar or dissimilar, and had to remember each of the letter sequences under a no-suppression control or a suppression condition. In the suppression condition, half of the subjects were engaged in an intermittent speech suppression and the other half were in an intermittent whistle suppression task. The phonological similarity effects appeared in the control condition, but not in the suppression condition, irrespective of the type of suppression. In the second experiment, the phonological similarity effect again disappeared in the intermittent whistling condition, but not in the condition in which the subjects required to engage a continuous whistling task. The results suggested that the effect of articulatory suppression was due to intermittent articulatory activity rather than the activation of an irrelevant phonology.
Our ability to recall a sequence of items is greatly affected by the phonological characteristics of the items. For instance, phonologically similar sequences of consonant letters lead to poorer serial recall than dissimilar ones. This is the so-called “phonological similarity effect” (Conrad & Hull, 1964), which indicates the importance of the role played by phonological coding in short-term memory.
Additional evidence of phonological coding has been accumulated in studies of articulatory suppression. In this method, the subject is required to repeatedly articulate some irrelevant speech sound such as the word “hiya” or “the.” The phonological similarity effect is abolished by articulatory suppression when the material is presented visually (Besner & Davelaar, 1982; Murray, 1968; Peterson & Johnson, 1971; Wilding & Mohindra, 1980). With auditory presentation however, the phonological similarity effect withstands articulatory suppression (Levy, 1971; Murray, 1968; Peterson & Johnson, 1971).
To account for these results, Baddeley (1990) introduced the idea of a phonological loop (originally, an articulatory loop; Baddeley, 1986; Baddeley & Hitch, 1974) as a subcomponent of working memory. The phonological loop comprises a phonological store and an articulatory control process. The phonological similarity effect is due to the operation of the phonological store. Auditory information has direct access to this store, but visual information only has access through the articulatory control process, which allows visually presented material to be phonologically coded. Articulatory suppression would simply serve to prevent visual information from entering phonological store. Thus, in visual presentation, when the operation of the articulatory control process is prevented by articulatory suppression, the phonological similarity effect disappears.
Articulatory suppression is a task which requires the subjects to utter a speech sound. Therefore, this activity includes at least the following components: The intention to speak, speech programming, actual articulation, and auditory feedback. Which component of articulatory suppression interferes with the articulatory control process? Recent research has indicated that attentional demands of speech (Saito, 1993b), actual articulation (Baddeley & Wilson, 1985; Saito, 1997), and auditory feedback (Gupta & MacWhinney, 1995; Saito, 1993c) are not major components of the articulatory suppression effects. Rather, the articulatory control process would depend on some form of speech motor programming or planning at a central level (Baddeley, 1990; Waters, Rochon, & Caplan, 1992), and the operation of that process must be disturbed by articulatory suppression acting on speech motor programming (Saito, 1993a, 1994, 1997).
For example, Saito (1997) showed that the phonological similarity effect appeared in the continuous “ah—” suppression condition despite the uttering of an irrelevant speech sound by the subjects. In contrast, the phonological similarity effect disappeared in an intermittent “ah, ah, ah …” suppression condition. The results suggested that the mere production of speech sound was not sufficient to suppress the activity of the phonological loop. There is a difference between the two articulatory suppression conditions in terms of the load on speech motor programming. Namely, continuous suppression may interfere less with speech motor programming than does intermittent suppression. For this reason, the articulatory control process can still operate in the continuous suppression condition, and the phonological similarity effect appears under this condition. But this is not the case in the intermittent condition.
Although Saito’s (1997) results can be explained in terms of description of speech motor programming, some researchers deny a direct link between speech motor programming and the phonological loop. Bishop and Robson (1989) observed the phonological similarity and the word-length effect in children with no articulate speech from birth, and it is difficult to accept that such children have developed an adequate speech motor program without ever having produced overt speech. Obviously, the representations in the phonological loop could not be articulatory gestures, according to Bishop and Robson (1989).
One hypothesis that can be generated is that the phonological representations in the phonological loop are more abstract ones, and of a kind available to individuals who have never produced speech. In other words, the representations are neither auditory nor articulatory, such that codes from a variety of sources are represented equivalently at an abstract phonological level (see Macken Sr Jones, 1995). With regard to speech motor programming, Bishop and Robson (1989) suggested that the abstract representations in normal individuals might be the ones which enter the system responsible for deriving a speech motor program but which do not themselves contain articulatory specifications. In other words, the speech motor programming may involve an activation and operation of these abstract phonological representations for speech production.
Under this view, there are two possible ways to account for the effect of articulatory suppression: (1) The effect is due to the repeated activation of the phonology. Namely, during the production of an irrelevant word, the phonological activation from an internal source (not from an auditory feedback) in the phonological store prevents the phonological loop from working. (2) Performing the articulatory suppression task disrupts the articulatory control process directly by its intermittent articulatory movement. Then, the phonological loop does not work well in that situation.
The irrelevant phonology hypothesis can explain Saito’s (1997) results as follows: There is a possibility that the continuous suppression task may differ from the intermittent suppression task in being less “phonological.” The continuous task is not properly phonological because there is not such a long “ah—” phoneme, at least in Japanese, which was the mother tongue of Saito’s subjects. This continuous task may produce a non-speech sound so that it can not activate a phonological representation internally, whereas the intermittent task repeatedly activates an irrelevant phonology. According to the articulatory movement hypothesis, however, there can be another explanation: There is a difference between the two articulatory suppression conditions in terms of the amount of articulatory movements. That is, continuous suppression, which requires subjects only to fix the positions of their teeth, tongue, and lips and to vibrate their vocal chords and which contains consequently small amount of articulatory movements, may interfere less with speech motor programming than does intermittent suppression, which contains greater movements of articulation.
Accordingly, both hypotheses can predict the fact that the phonological loop still operates in the continuous suppression condition but not in the intermittent condition. Thus, a question to be addressed is whether the effect of articulatory suppression is due to the activation of an irrelevant phonology or to intermittent articulatory movements. To answer this question, the current experiments used a whistle task as an articulatory suppression task. That task requires the subjects to perform an intermittent articulatory activity without activation of the irrelevant phonology for speech production. The effects of intermittent whistling on the phonological similarity effects were examined.
Experiment 1 Two types of articulatory suppression were used, i.e., an intermittent speech (ah, ah, ah…) suppression and an intermittent whistle (pip, pip, pip…) suppression. Subjects were tested for the serial recall of the visual letter sequences that were phonologically either similar or dissimilar. If the irrelevant phonology hypothesis is correct, the phonological similarity effect would be observed in the whistle suppression condition but not in the speech suppression. On the other hand, if the articulatory movement hypothesis is right, the phonological similarity effect would appear in neither the whistle nor the speech suppression conditions.
METHOD Participants. Twenty-four undergraduates, 4 men and 20 women, at Osaka University of Education participated. Their average age was 19.9 years. The participants were non-native speakers of English, but were familiar with the alphabet letters used in this experiment as the to-be-remembered materials. Design. A three-way factorial design (type of activity x articulation x phonological similarity) was used. Type of activity (speech versus whistle) was manipulated as a between-subjects factor. The factors of articulation (no suppression control versus suppression) and phonological similarity (dissimilar versus similar) were manipulated within subjects.
All participants took part in an experimental session composed of four blocks; Dissimilar-Control (DC), Dissimilar-Suppression (DS), Similar-Control (Sc), and Similar-Suppression (ss). Each block had two unscored practice trials followed by five test trials. The orders of the blocks were as follows; DC-SC-DS-SS, SC-DC-SS-DS, DS-SS-DC-SC, and SS-DS-SC-DC. Each of these orders was presented to and performed by six participants. In the suppression conditions (DS or SS), half of the participants were engaged in a speech suppression task and the remaining participants were in a whistle suppression task. Materials and Procedure. To eliminate the effects of auditory feedback, the participants were exposed to an irrelevant speech sound, “ah, ah, ah…,” during the serial memory task in all conditions. Although auditory feedback is not major component of the articulatory suppression effect (Gupta & MacWhinney, 1995; Saito, 1993c), it has some impact on immediate serial recall performance (Gupta & MacWhinney, 1995). Further, it is known that a repeated presentation of a vowel produced the impairment of recall in some situations (LeCompte, 1995, Experiment 4; but see Jones, Madden, & Miles, 1992). It may be possible that the auditory feedback of a suppression task has some effects on the memory performance in this experiment. Since the two types of suppression task differ in auditory feedback (speech or non-speech feedback), one type of feedback sound may affect the memory performance differently from another. The procedure of a repeated presentation of the “ah” sound was expected to have a consequence. That is, the irrelevant speech sound should equalize an external auditory interference in one condition with that in the another condition, despite the difference in auditory feedback between speech and whistle activity.
The task requirement in the no suppression control condition was to perform the memory task only. In the suppression conditions, the memory task had to be performed with either speech suppression or whistling suppression.
Memory task. On each trial, subjects were shown a sequence of six letters, either B, C, D, G, P, T (phonologically similar set) or K, J, M, Q, R, Y (dissimilar set), in a different order each time. The task of the subjects was to remember the letters in the correct order. Following an auditory warning signal, letters were presented sequentially for 500 ms each on a CRT controlled by a Macintosh (Performa 5220) via HyperCard software. A 5-s unfilled delay followed the termination of the last letter. The end of the delay was indicated by a question mark and an auditory “beep” signaling the start of a recall period. Written serial recall was used. The subjects wrote an answer from left to right on a recall sheet which contained six blank boxes. They were allowed to retrace leftward to change their answer but they were not allowed to fill in a blank on the right side until all of the preceding blanks to the left were filled. No blanks were allowed to remain unfilled. When the subjects finished their recall, the recall sheet was replaced by a new one for the following trial. In all conditions, subjects wore headphones.
Irrelevant speech. The auditory material was a repeating “ah” produced by a male voice, and was presented from the “‘” signal that indicated the start of a trial until the “?” signal that indicated the start of a recall period. The rate of the repetition was approximately three to four times per second. It was presented via stereo dynamic headphones (AIWA; HP-x35) from a recording of the sounds held in digital format within the Macintosh.
Suppression. Subjects of the speech suppression group were required to repeat “ah” in synchronization with the presentation of the irrelevant speech sound. For the whistle group, subjects were required to whistle intermittently in synchronization with the presentation of the irrelevant speech sound. They were instructed to make a clear sound by forcing air through a narrow hole formed by their lips. If the tone produced by the subjects was not a high clear one, it was stressed that all they had to do was to shape their lips and to breathe out intermittently.
RESULTS Letters were scored as correct if they were recalled in the same serial position as they were presented. Table 1 shows the proportion of correct responses in the various experimental conditions. Considering the speech suppression group, a clear phonological similarity effect was obtained in the control condition but not in the suppression condition. It is also evident from Table 1 that, for the whistle suppression group, a phonological similarity effect was obtained in the control condition but not in the suppression condition. A three-way (ANOVA) analysis of variance, with the variables type of activity, phonological similarity, and articulation, demonstrated reliable main effects of both phonological similarity, F (1, 22) = 16.64, MS, = .012, p
A one-way ANOVA for dissimilar lists revealed a main effect of suppression, F (1, 22) = 20.24, MSe = .023, p
DISCUSSION Overall, the results for the speech and whistle suppression groups fell into a consistent pattern. That is, the phonological similarity effects appeared in the no suppression control but not in the suppression condition, irrespective of the type of articulatory activity.
Although the results of Experiment 1 may provide evidence that the activation of irrelevant phonology is not essential to the articulatory suppression effect, they do not on their own provide evidence for supporting the articulatory movement hypothesis. This is because a continuous suppression task, which was used by Saito (1997), was not included in the experimental design. Namely, the amount of articulatory activity was not manipulated directly in the first experiment.
Experiment 2 The purposes of this experiment are (1) to replicate a part of the results of Experiment 1, that is, to show the disappearance of phonological similarity effect under the condition of an intermittent whistle, and (2) to test the articulatory movement hypothesis by directly comparing an intermittent whistle and a continuous whistle condition. Because the continuous whistling task would demand the minimum movements of articulatory organs, it is not expected to interfere with speech motor programming. This contrasts with the intermittent whistling task, in which more articulatory movements are required. If the articulatory movement hypothesis is correct, the phonological similarity effect would be observed in the continuous whistle condition but not in the intermittent whistle condition.
METHOD Participants. Sixteen undergraduates, 5 men and 11 women, at Osaka University of Education participated. Their average age was 19.6 years. The subjects were non-native speakers of English, but they were familiar with the alphabet characters.
Design. A two-way factorial design (type of suppression x phonological similarity) was used. Type of suppression (intermittent versus continuous) and phonological similarity (dissimilar versus similar) were manipulated within subjects.
All participants took part in an experimental session composed of four blocks; Dissimilar-Intermittent (DI), Dissimilar-Continuous (DC), Similar-Intermittent (SI), and Similar-Continuous (Sc). Each block included two unscored practice trials followed by five test trials. The orders of the blocks were as follows; DI-SI-DC-SC, SI-DI-SC-DC, DC-SC-DI-SI, and SC-DC-SI-DI. Each of these orders was presented to and performed by four participants. Materials and Procedure. The general procedural details for this experiment were similar to those in Experiment 1, except for the following points: (1) All participants were required to engage whistle suppression tasks. (2) They engaged two types of whistle suppression task. (3) They were not exposed to the irrelevant speech sound and did not wear headphones on any trials.
Memory task. The procedure for memory task in this experiment was the same as that of Experiment 1. Whistle tasks. Participants were required to whistle either in a continuous or an intermittent ways from the “‘”” signal that indicated the start of a trial until the “?” signal that indicated the start of a recall period. Under the condition of continuous whistle, subjects were encouraged to produce a clear tone continuously (pih–) by forcing air, without break, through a narrow hole formed by their lips. They were allowed to breathe in once or twice per trial. In the intermittent condition, they were encouraged to repeat whistling in a staccato fashion (pip, pip, pip…) at a rate of approximately three or four times per second and were cautioned if their rate of whistling showed signs of becoming slower. RESULTS Letters were scored as correct if they were recalled in the same serial position as they were presented. Table 2 shows the proportion of correct responses in the various experimental conditions. It is evident from the table that a phonological similarity effect was obtained in the continuous suppression condition but not in the intermittent suppression condition. A two-way ANOVA, with type of suppression and phonological similarity, demonstrated a reliable main effect of phonological similarity, F (1, 15) = 12.68, MSe = .019, p
Subsidiary analyses further probed the interaction between suppression and phonological similarity. A one-way analysis indicated an advantage for the dissimilar sequence over the similar one in the continuous suppression condition, F (1, 15) = 22.20, MSe = .014, p
A one-way ANOVA for dissimilar lists led to a main effect of suppression, F (1, 15) = 5.11, MSe = .030, p
DISCUSSION As predicted from the articulatory movement hypothesis, recall performance for the dissimilar lists was better than that for the similar lists only in the continuous whistle condition. That is, the phonological similarity effect was observed in the continuous whistle condition and not in the intermittent whistle condition. It seems reasonable to conclude that performing an articulatory suppression task disrupts the articulatory control process by its intermittent articulatory activity. General Discussion In both experiments, the phonological similarity effects disappeared under the conditions in which the subjects were required to engage an intermittent whistle task. As mentioned in the Introduction, the phonological similarity effect reflects a function of the phonological store. Visual material has access to this store via the articulatory control process, the operation of which would depend on some form of speech motor programming or planning at a central level (Baddeley, 1990; Saito, 1993a, 1994, 1997; Waters, Rochon, & Caplan, 1992). The intermittent suppression task demands the articulatory movements by the subjects and is expected to interfere with the speech motor programming. Because the operation of the articulatory control process is prevented by the intermittent articulatory activity, phonological information concerning the stimuli could not enter the phonological store. Thus, the phonological similarity effect disappears in the intermittent suppression conditions. In contrast, the continuous whistle task would demand the minimum of movements on articulatory organs and would not be expected to interfere with speech motor programming. For this reason, the phonological loop works in the continuous whistling condition.
The results of the present experiments suggest that articulatory suppression might affect memory performance by its intermittent articulatory activity rather than by the activation of an irrelevant phonology. The idea of articulatory movement account may also prove useful in considering related data. For example, Conrad (1970) reported that despite never having being able to hear, congenitally deaf children who can orally communicate with others showed phonological confusions in remembering consonant sequences. This result shows the importance of articulatory coding in short-term memory.
What has to be noticed is that an intermittent “non-speech” whistling could act as an articulatory suppression task. This result indicates the importance of the intermittent activity on the effects of articulatory suppression. This idea, however, immediately raises the question of whether the effect of articulatory suppression is due to the intermittent “articulation” or the intermittent activity itself. Judging from the previous findings (Baddeley, Eldridge, & Lewis, 1981; Logie & Baddeley, 1987; Saito, 1993a, 1997), the “articulation” would be important for the effects. In fact, a simple finger tapping in synchronization with an irrelevant speech sound (Saito, 1997) or with a computer-produced tone (Saito, 1993a) did not decrease the performance of serial recall and did not affect the magnitude of the phonological similarity effect. In addition, a tapping task at the same rate as articulatory suppression did not reduce the performance of counting (Logie & Baddeley, 1987) and of reading (Baddeley, Eldridge, & Lewis, 1981) as much as articulatory suppression did. It is obvious that finger tapping in synchronization with simple rhythm can not disrupt the activity of the phonological loop. On the other hand, a repeated production of a single vowel “ah” is sufficient to abolish the phonological similarity effect (Experiment 1), as Murray (1968) showed with a single syllable “the.” The intermittent activity by the articulatory organs would be crucial for the occurrence of the articulatory suppression effects.
It is concluded that the concurrent articulation suppresses the phonological loop by its intermittent articulation, neither by the activation of an irrelevant phonology, nor by mere occupation of the articulatory organs, nor by the intermittent activity per se. I would like to thank Colin M. MacLeod, Catherine G. Penney, Murray Singer, and an anonymous reviewer for their valuable comments on earlier versions of this paper. Correspondence should be addressed to: Satoru Saito, Department of Psychology, Osaka University of Education, 4-698-1 Asahigaoka, Kashiwara, Osaka 582-8582 Japan (e-mail: saitoXcc.osaka-kyoiku.ac.jp).
Baddeley, A. D. (1986). Working memory. Oxford: Oxford University Press.
Baddeley, A. D. (1990). Human memory: Theory and practice.
Hove: Lawrence Erlbaum Associates.
Baddeley, A. D., Eldridge, M., & Lewis, V. (1981). The role of subvocalization in reading. Quarterly Journal of Experimental Psychology, 33, 439-454.
Baddeley, A. D., & Hitch, G. J. (1974). Working memory. In G. Bower (Ed.), The psychology of learning and motivation V71z, (pp.47-90). New York: Academic Press. Baddeley, A. D., & Wilson, B. A. (1985). Phonological coding and short-term memory in patients without speech. Journal of Memory and Language, 24, 490-502.
Besner, D., & Davelaar, E. (1982). Basic processes in reading: Two phonological code. Canadian Journal of Psychology, 36, 701-711.
Bishop, D. V. M., & Robson, J. (1989). Unimpaired short-term memory and rhyme judgment in congenitally speechless individuals: Implications for the notion of “articulatory coding”. Quarterly Journal of Experimental Psychology, 41A, 123-140.
Conrad, R. (1970). Short-term memory processes in the deaf.
British Journal of Psychology, 61, 179-195. Conrad, R., & Hull, A. J. (1964). Information, acoustic confusion, and memory span. British Journal of Psychology, 55, 429-432.
Gupta, P., & MacWhinney, B. (1995). Is the articulatory loop
articulatory or auditory? Reexamining the effects of concurrent articulation on immediate serial recall. Journal of Memory and Language, 34, 63-88.
Jones, D. M., Madden, C., & Miles, C. (1992). Privileged access by irrelevant speech to short-term memory: The role of changing state. Quarterly Journal of Experimental Psychology, 44, 645-669.
LeCompte, D. C. (1995). An irrelevant speech effect with repeated and continuous background speech. Psychonomic Bulletin & Review, 2, 391-397.
Levy, B. A. (1971). Role of articulation in auditory and visual short-term memory. Journal of Verbal Learning & Verbal Behavior, 10, 123-132.
Logie, R. H., & Baddeley, A. D. (1987). Cognitive processes in counting. Journal of Experimental Psychology: Learning Memory, and Cognition, 13, 310-326.
Macken, W. J., & Jones, D. M. (1995). Functional characteristics of the inner voice and the inner ear: Single or double agency? Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 436-448.
Murray, D. J. (1968). Articulation and acoustic confusability in short-term memory. Journal of Experimental Psychology, 78, 679-84.
Peterson, L. R., & Johnson, S. T. (1971). Some effects of minimizing articulation on short-term retention. Journal of Verbal Learning and Verbal Behavior, 10, 346-54. Saito, S. (1993a). The disappearance of the phonological similarity effect by complex rhythmic tapping. Psychologia, 36, 27-33.
Saito, S. (1993b). Influence of articulatory suppression and memory updating on phonological similarity effect. Japanese Journal of Psychology, 64, 289-295. (In Japanese with English abstract).
Saito, S. (1993c) Phonological similarity effect is abolished by a
silent mouthing task. Perceptual and Motor Skills, 76, 427-431. Saito, S. (1994). What effect can rhythmic finger tapping have on the phonological similarity effect? Memory & Cognition, 22, 181-187.
Saito, S. (1997). When articulatory suppression does not suppress the activity of the phonological loop. British Journal of Psychology, 88, 565-578.
Waters, G. S., Rochon, E., & Caplan, D. (1992). The role of high-level speech planning in rehearsal: Evidence from patients with apraxia of speech. Journal of Memory and Language, 31, 54-73.
Wilding, J., & Mohindra, N. (1980). Effects of subvocal suppression, articulating aloud, and noise on sequence recall. British Journal of Psychology, 71, 247-261.
Date of acceptance: October 20, 1997
Copyright Canadian Psychological Association Mar 1998
Provided by ProQuest Information and Learning Company. All rights Reserved