1 Applied Psycholinguistics 24 (2003), 173–193 Printed in the United States of America DOI: 10.1017.S0142716403000092 Bilingual and monolingual processing of competing lexical items VIORICA MARIAN Northwestern University MICHAEL SPIVEY Cornell University ADDRESS FOR CORRESPONDENCE Dr. Viorica Marian, Department of Communication Sciences and Disorders, Northwestern University, 2240 Campus Drive, Evanston, IL 60208-3570. E-mail: [email protected] ABSTRACT Performance of bilingual Russian–English speakers and monolingual English speakers during audi- tory processing of competing lexical items was examined using eye tracking. Results revealed that both bilinguals and monolinguals experienced competition from English lexical items overlapping phonetically with an English target item (e.g., spear and speaker ). However, only bilingual speakers experienced competition from Russian competitor items overlapping crosslinguistically with an En- glish target (e.g., spear and spichki, Russian for matches ). English monolinguals treated the Russian competitors as they did any other filler items. This difference in performance between bilinguals and monolinguals tested with exactly the same sets of stimuli suggests that eye movements to a crosslinguistic competitor are due to activation of the other language and to between-language com- petition rather than being an artifact of stimulus selection or experimental design. With the majority of the world’s population speaking more than one language (Romaine, 1995), studying bilingualism and multilingualism can provide valu- able insights into human cognition and language. The capability of one cognitive system to successfully manage two languages is striking. Do bilinguals use the two languages independently, alternating between them by turning them on and off, or do they constantly keep both languages active and process the two in parallel at all times? The traditional language switch hypothesis, according to which bilinguals are able to selectively activate and deactivate their two lan- guages (Gerard & Scarborough, 1989; MacNamara & Kushnir, 1971), has been challenged by a number of recent findings. Parallel activation has been inferred from early studies using the bilingual Stroop task (Chen & Ho, 1986; Preston & Lambert, 1969), code switching (Grainger, 1993; Grainger & Dijkstra, 1992; Li, 1996; Soares & Grosjean, 1984), interlingual homographs (Dijkstra, Timmer- mans, & Schriefers, 1997; Dijkstra, van Jaarsveld, & ten Brinke, 1998), cog- 2003 Cambridge University Press 0142-7164/03 $12.00
2 Applied Psycholinguistics 24:2 174 Marian & Spivey: Bilingual and monolingual processing nates (DeGroot & Nas, 1991; Kroll & Stewart, 1994), and phoneme monitoring (Colome, 2001). Recent evidence for parallel activation comes from visual word recognition tasks that test masked orthographic priming (Bijeljac–Babic, Biar- deau, & Grainger, 1997), interlingual neighbors (van Heuven, Dijkstra, & Grain- ger, 1998), and phonological overlap (Brysbaert, Van Dyck, & Van de Poel, 1999; DeGroot, Delmaar, & Lupker, 2000; Dijkstra, Grainger, & van Heuven, 1999). Attempts to extend the hypothesis of generalized lexical access from the orthographic to the phonological domain using code-switching were also encouraging (Doctor & Klein, 1992; Nas, 1983). In the auditory domain, some of the most compelling evidence supporting automatic activation of both lexicons during monolingual input comes from re- search investigating spoken language processing in bilinguals using eye tracking (Marian, 2000; Spivey & Marian, 1999). The eye-tracking technique, merging input from both the visual and auditory modalities, allows one to index the activation of a second language nonlinguistically. It allows testing the process- ing of both languages without compromising a monolingual language set, some- thing that is otherwise not possible with spoken language. In the eye-tracking methodology, participants are given spoken instructions to move objects around and their eye movements to the various objects are recorded. Although they rarely pick up incorrect objects, it is often observed that participants briefly fixate objects that have similar phonology to the spoken word. In fact, Allo- penna, Magnuson, and Tanenhaus (1998) showed that the probabilities of eye movements to objects with phonological similarity closely match the activation curves of the TRACE model of speech perception (Elman & McClelland, 1986; McClelland & Elman, 1986), thus providing the linking hypothesis between eye movements and lexical activation. Using this methodology, Spivey and Marian (1999) presented Russian– English bilinguals with a visual display consisting of four objects and asked them to manipulate a target object. The onset of the name of the target object bore phonetic similarity to the name of one of the other objects in the other language. For example, when instructed, “Poloji nije krestika” ‘Put the marku stamp below the cross’, the visual display in front of the subject contained, among other objects, a the name of which shares several phonemes marker, marka, with the Russian word for stamp. It was found that while processing the target word marka , Russian–English bilinguals made eye movements to the between-language competitor word marker , thus suggesting that lexical items in both languages were activated simultaneously, even though only one language was being used. Although this work shows crosslinguistic activation in bilingual spoken lan- guage processing, it leaves many questions unanswered. It does not, for exam- ple, indicate what happens during bilingual language processing when competi- tion takes place in more complex visual content. In everyday environments, while processing spoken language, bilinguals are surrounded by a multitude of objects. Some of them do indeed compete crosslinguistically, others compete within the same language, and in many cases there may be simultaneous compe- tition from both languages. What happens under these circumstances of simulta- neous competition from both languages?
3 Applied Psycholinguistics 24:2 175 Marian & Spivey: Bilingual and monolingual processing A second question arises from the fact that Spivey and Marian (1999) found an asymmetry in their results, with significant competition from the second lan- guage into the first language, but not from the first language into the second. We hypothesized that the asymmetry may have been a result of the increased activation of English among the tested bilinguals due to their immersion in an English-speaking environment, as well as possible differences across languages in the stimuli selected on variables such as phonetic overlap and word frequen- cies. In the present work, we aimed to further study the between-language com- petition phenomenon observed by Spivey and Marian (1999) while trying to increase the activation of the Russian language and control for amount of pho- netic overlap and word frequencies in the two languages. We predicted that between-language competition would be observed from both languages and into both languages. In addition, we wanted to see if bilinguals would show within-language com- petition from items whose name bore phonetic overlap in the same language, as did English monolinguals (Allopenna et al., 1998; Tanenhaus, Spivey–Knowl- ton, Eberhard, & Sedivy, 1995), thus extending the within-language competition phenomenon to the bilingual domain. Therefore, we had three goals in studying bilinguals in the present work: (a) to replicate between-language competition from the second language into the first language and examine whether competi- tion from the first into the second language could take place, (b) to examine within-language competition in bilinguals, and (c) to examine language process- ing during simultaneous competition from both languages. In addition to bilinguals, a second experiment with monolingual English speakers was run. For this experiment, the same stimuli, setup, procedure and design were used as with bilinguals, but testing was done in English only. The goals of the monolingual experiment were to examine whether, during English trials, monolingual English speakers would experience between-language com- petition from Russian competitors, within-language competition from English competitors, and/or simultaneous competition from both languages. Performance of the monolingual speakers could then be directly compared to that of the bilingual group on the English part of the experiment. If both bilinguals and monolinguals show competition from English but only bilinguals also show competition from Russian, then it can be concluded that eye movements to phonetically overlapping items are indeed a result of parallel activation of both languages and of between-language competition, rather than an artifact of stimu- lus selection or study design. If, on the other hand, both groups perform equally in all conditions (showing competition either from both Russian and English or from neither), then the eye-tracking paradigm is ill suited for studying language activation and processing in bilinguals. Testing these hypotheses was the major objective of the monolingual–bilingual comparison. EXPERIMENT 1 In Experiment 1, Russian–English bilingual speakers were tested in their first and second languages to examine spoken language processing under three con- ditions: (a) when lexical items competed for activation with a target item be-
4 Applied Psycholinguistics 24:2 176 Marian & Spivey: Bilingual and monolingual processing tween-languages, (b) when lexical items competed for activation with a target item within-language, and (c) when lexical items competed for activation with a target item simultaneously from both between and within languages. Methods Fifteen Russian–English bilinguals participated in the study (5 Participants. males, 10 females). All were students at Cornell University, and their mean age was 22.04 years ( SD = 3.77). All were Russian–English bilinguals who were born in the former Soviet Union and immigrated to the United States at a mean SD age of 15.62 years ( 3.65). All participants were fluent in both languages = and had obtained sufficiently high scores on standardized SAT tests to be admit- ted to Cornell; none were enrolled in English as a Second Language classes. When asked what language they used most, participants reported that Russian was the language they used most until the average age of 18 years (i.e., for another 2–2.5 years after immigration), at which point English became the lan- guage they used most and remained so for the past 4 years on average. In addition, participants were asked to self-report their language preference. Six participants indicated that Russian was their preferred language, five indicated that English was their preferred language, and four reported no language prefer- ence. It is also worth mentioning that this group of bilinguals was similar to that tested by Spivey and Marian (1999). In that study, participants were 7 male and 5 female Russian–English bilinguals with a mean age of 20.35 years at the time of the experiment and 14.01 years at the time of immigration. All participants were paid for their participation. At the end of the experiment, participants were asked whether they noticed anything unusual and what they thought the objectives of the study were. Normally, in this task any participant who mentions noticing similar sounding object names is excluded from data analy- ses. In this experiment, none of the participants reported noticing any similarities. Apparatus. A headband-mounted ISCAN eyetracker was used to record partici- pants’ eye movements during the experiment. A scene camera, yoked with the view of the tracked eye, provided an image of the participant’s field of view. A second camera, which provided an infrared image of the left eye, allowed the software to track the center of the pupil and the corneal reflection. Gaze position was indicated by crosshairs superimposed over the image generated by the scene camera. The output was recorded onto a Hi8 VCR with frame by frame play- back. All the instructions were prerecorded by a fluent Russian–English bilingual speaker who acquired both languages in early childhood and had no noticeable accent in either language. The speech files were recorded and played on a Mac- intosh computer, and the audio record was synchronized with the video record for data analysis. Design. All participants were tested in two parts, a Russian part and an English part, with order counterbalanced across subjects. Each part consisted of 20 trials,
5 Applied Psycholinguistics 24:2 177 Marian & Spivey: Bilingual and monolingual processing which were equally distributed across four conditions. In the no-competition condition, of the four objects presented in the display, one was the target object and three were control filler objects. The target object was the object actively named in the experiment. The filler objects were objects whose names did not overlap with the name of the target object in either language. This first condition served as the baseline for all analyses. In the between-language competition condition, of the four objects presented in the display, one was the target object, one was the between-language competi- tor, and two were filler objects. The between-language competitor was an object whose name in the other language had phonetic overlap with the name of the speaker was the target target object. For example, during the English part, if spichki object, then matches ) was the between-language competitor object. The ( name of the between-language competitor was never spoken in either language during the entire experiment. In the within-language competitor condition, of the four objects presented in the display, one was the target object, one was the within-language competitor, and two were filler objects. The within-language competitor was the object whose name carried phonetic similarity to the target object in the same lan- guage. For example, during the English part, if the target object was a speaker, then the within-language competitor was a spear . Similarly, in the Russian part, if the target object was spichki , then the within-language competitor was spitsy ( knitting needles ). The name of the within-language competitor was also never spoken during the entire experiment. Finally, in the fourth condition, of the four objects presented in the display, one was a target object, one was a between-language competitor, one was a within-language competitor, and one was a filler object. This fourth condition allowed testing a situation in which simultaneous between-language and within- language competition takes place. The four conditions were intermixed through- out the experiment. For a given target item, across all four of its conditions (between-language competitor present/absent by within-language competitor present/absent), the various competitor and filler objects had fixed locations in which they would be presented. This allowed us to compare eye movements to competitors and fillers by examining fixations of the same location within a display. Stimuli. complete list of all target items, between-language competitors, A within-language competitors, and fillers for both Russian and English can be found in the appendices. To avoid potential confounds, in each trial we considered such variables as the physical similarity of the items, the word frequencies in the two languages, and the amount of phonetic overlap. During the experiment, similarities in the physical properties (size, shape, color) of a target object and one of the filler objects in the display were noticed for one Russian set (a balloon and a pear) and one English set (a greeting card and a napkin). As a result, all trials contain- ing these sets were discarded from analyses and are not considered in the results and discussion. To compute word frequency, we used three sources. For the English language
6 Applied Psycholinguistics 24:2 178 Marian & Spivey: Bilingual and monolingual processing we used Zeno et al.’s (1995) Word Frequency Guide based on a corpus of 17,274,580 word tokens. For the Russian words we used Lenngren’s (1993) frequency dictionary based on a corpus of 1,000,000 word tokens, as well as Zasorina’s (1977) frequency dictionary based on 40,000 word tokens. In addi- tion, we translated all the Russian words and considered the frequency of the translated words in the English language and translated all the English words into Russian and considered the frequency of the translated words in the Russian language based on the two Russian sources. Both raw word frequency and word frequency adjusted for dispersion were considered. None of the performed anal- yses showed any statistically significant difference in the frequency of the target and competitor items in either language. Appendices A and B provide the word frequencies of target, competitor, and filler stimulus items that were used and their translation equivalents. Phonetic overlap was analyzed based on the raw number of overlapping pho- nemes and the proportion of word overlap across two items. We performed T tests across the two languages and for each language separately. Overlap within and between languages was considered. No significant difference was found in any of the analyses. Tables 1 and 2 provide phonetic transcriptions of stimuli used in the two experiments and amount of phonetic overlap between target and competitor items, following International Phonetic Alphabet transcriptions. Procedure. Upon arrival to the lab, subjects were greeted and interacted with exclusively in the language appropriate for that part of the experiment. In addi- tion, efforts were made to increase the activation of Russian, because Russian was a passive language in the subjects’ overall environment at the time. These efforts included providing all the instructions in Russian prior to the Russian part of the experiment, having subjects sign a Russian-language consent form before the Russian part, and playing popular Russian songs via a tape-recorder at the time of arrival to the lab or before the beginning of the Russian part. Participants were then seated at arm’s length from a 61 by 61 cm white board set on a table. The board was divided into nine equal squares, and a black cross in the center square served as a neutral fixation point. After the eyetracker was calibrated, each participant was presented with 40 different displays of object combinations (20 displays in each language). For each display the subjects were asked to look at the central cross, followed by instructions to manipulate the objects in the display. Examples of verbatim instructions are “Pick up the speaker. Put the speaker below the cross” in English and “Podnimite spichki. Polojite spichki nije krestika” (“Pick up the matches. Put the matches below the cross”) in Russian. For each instruction to manipulate an object from a phonetically overlapping set, there were two filler instructions to manipulate a filler object that did not overlap phonetically with any other item in the display. These filler trials served two purposes. First, they were used to prevent subjects from noticing the overlap and guessing the hypothesis of the study. Second, because the overlapping com- petitor objects and their corresponding distractor objects all served as filler ob- jects in these filler instructions, we were able to perform control comparisons
7 b 1 2 3 3 3 1 2.20 0.92 No. of at Onset Phonemes Overlapping lz] t] ]2 E ε 7 b R ss ]1 + α bst ]3 ∫ε vl m] k] R v k] m] R v ]3 v ∫ ν R v α Item English Competitor α ovel [ arbles [m ark [b Spear [spi Book [b Plum [pl Lobster [l Chess set [t Gum [g c c a,b a a a No. of at Onset Phonemes Overlapping 1.80 2.20 0.63 0.42 ]1 gtei α an j ]1 ]2Car[k E kdl E h k α ∫ α [l ]2B p ]3 t ε rto ]2M j ]2 E ]2 Item r ki E Russian Competitor ε ]2Sh α ]1 ∫ n ε j rh tj k E ε ∫ rk j [ t rik α [ jk α α spit α pl ∫α [ b m [ [ [ bub g [ [ [ EB4R84 '"68" ;"D8" #"DN"H 7"[email protected]" A:"H\, #J$,> S"D48 9"8 *:b >@(H,6 Q,D,B"N" ] R bd wai R ] ] α 7 7 k R α k] ] d] R R g] R k] n] e v ∫ α ∫α v v Phonetic transcriptions of stimuli used with the English target and amount of phonetic overlap 5. Marker [m 1. Speaker [spik 6. Barbed wire [b 9. Card [k 8. Gun [g 7. Plug [pl 2. Boot [but] 3. Shark [ 4. Chair [t The means and standard deviations include the vowels that have strong similarities across languages and are likely Indicates affricate overlap. Indicates strong similarities in vowel quality following the overlap. Table 1. between target and competitor items following International Phonetic Alphabet transcriptions c a b to sound the same to a late bilingual. Mean English Target Standard deviation 10. Lock [l
8 b 2.60 0.70 No. of at Onset Phonemes Overlapping ]4 E k ∫ ]2 ]3 ]2 k ]2 t rto E ]3 E E ]2 ]2 ]3 i ak α j E E n j lst rh k ε ]3 j [ rv pat c α pk ε α ∫ g α α t spits- [ b ∫α Item [ [ [l [ [ Russian Competitor bub pl [marl [ [ EB4P\y 7"[email protected]" '":FHJ8 A:"V [email protected]"H" #J$,> Q,D&b8 ;"D:b S"B8" #"DN"H c c a,b c a a a 1 1 2 1 2 No. of at Onset Phonemes Overlapping 1.70 2.20 0.67 0.42 ]2 R bd wai R ]2 α 7 k R ]3 Item α k] ]1 English Competitor m] R R R k] ]2 k] v n] e ν R ∫ v ∫α v α [g [k arker [m hair [t ook [b arbed wire [b ock [l pear [spi hark [ lum [pl B ]L gtei α an j ]C E kdl h α α [l p ]S ε ]M ]P j ]B r ki E ε ]S ]Gun ∫ E ]Car ε tj j E ∫ E rk t α rik nk [ rt jk α i] spit pl α α α ∫α [ [ m [ b k g [ [ [ [ [bus- Phonetic transcriptions of stimuli used with the Russian target and amount of phonetic overlap 9"8 *:b >@(H,6 ;"D8" EB4R84 #JFZ 7"DH" '"68" S"D48 #">8" A:"H[, Q,D,B"N" 1. 5. 2. 9. 8. 3. 6. 7. 4. The means and standard deviations include the vowels that have strong similarities across languages and are likely Indicates affricate overlap. Indicates strong similarities in vowel quality following the overlap. between target and competitor items, following International Phonetic Alphabet transcriptions Table 2. c a b to sound the same to a late bilingual. 10. Russian Target Mean Standard deviation
9 Applied Psycholinguistics 24:2 181 Marian & Spivey: Bilingual and monolingual processing and ensure that any observed effects were not a by-product of physical differ- ences between the competitor object and its corresponding filler. Analyses. Recording of eye movements provided two measures, proportion and pattern. The pattern of eye movements is provided for descriptive purposes only; all inferential statistics were performed on proportions of eye movements. The proportion of eye movements to the competitor item in a competition condi- tion was compared to the proportion of eye movements to a nonoverlapping filler item in the same location in the control condition. This allowed us to tabulate the proportion of trials in which the participants looked at the same display square across conditions, with that square containing either a control filler object or a competitor object. Three-way analyses of variance (ANOVAs) were performed for between-language and within-language competition, using language (Russian vs. English), between-language competition (present vs. ab- sent), and within-language competition (present vs. absent). These analyses were performed across all four conditions: for example, if both the between-language competition variable and the within-language competition variable indicated ab- sent, then that referred to the control condition; if only one indicated present, this referred to competition from one language only; and if both indicated pres- ent, then that referred to the simultaneous competition condition. Because the between-language competitor and the within-language competitor were neces- sarily in different locations in the display (and were present simultaneously in one of the conditions) and because they had different objects as their corre- sponding control fillers, we conducted separate three-way repeated-measures ANOVAs for eye movements to the between-language competitor location and the within-language competitor location. Results Proportion of eye movements. The proportion of eye movements to competitor items was compared to the proportion of eye movements to filler items in the same location. Three-way ANOVAs (language, presence or absence of between- language competitor, and presence or absence of within-language competitor) were computed by subjects and by items for between-language competition (i.e., looks to the display square that contained either the between-language competi- tor or the control filler object). Results revealed a main effect of presence or absence of the between-language competitor. Russian–English bilinguals looked at the between-language competitor in 15% of all trials and to the control filler in the same location in 6% of all trials. This difference was significant by both subjects and items, F 1 (1, 14) = 7.40, p < .05; F 2 (1, 16) = 5.24, p < .05. No effects of language (Russian vs. English) or presence/absence of the within- language competitor were observed, suggesting that bilinguals experienced between-language competition in both languages during between-language com- petition trials and simultaneous competition trials. There was no interaction be- tween any of the variables. Asim as performed for within-language competition. × 2 × 2ANOVAw ilar 2
10 Applied Psycholinguistics 24:2 182 Marian & Spivey: Bilingual and monolingual processing The results suggested a main effect of presence or absence of the within-language competitor but no effect of language or presence/absence of the between-language mp eti to r. Fo r wit hi n-l an gua ge com pe tit io n, Ru s sia n– Eng co sh bilinguals looked li at the within-language competitor in 20% of all trials and to the control filler in the same location in 10% of all trials. This difference was significant both by F 1(1, 14) = 1 0.05, p < subjects and by items, F 2(1,16) = 8.06, p < .05. No .01; interaction between variables was observed. That is, bilinguals experienced within- language competition in both their first language and their second language during within-language competition trials and simultaneous competition trials. A comparison of the first and second halves of the experiment in order to test for order effects failed to reveal a significant difference in patterns of competi- tion. Within-language competition, between-language competition, and simulta- neous within- and between-language competition were not significantly different in the two parts. To get a sense of the time course of fixations, the Eye-movement patterns. pattern of bilinguals’ eye movements was examined. When only the between- language competitor was present and participants fixated it, they did so on at average of 475 ms after the onset of the target word. On those trials, they then fixated the target object an average of 817 ms after the onset of the target word. When only the within-language competitor was present and participants fixated it, they did so 602 ms after the onset of the target word. On those trials, they looked at the target object an average of 808 ms after the onset of the target word. In the simultaneous within- and between-language competition condition, of the times when there was a look to a competitor object, 56% of the time the look was to a within-language competitor only, 32% of the time the look was to a between-language competitor only, and 12% of the time participants looked at both within- and between-language competitors. On those few trials when participants looked at both within- and between-language competitors, 50% of the time they looked at the within-language competitor first and 50% of the time they looked at the between-language competitor first. In the simultaneous competition condition, participants looked at the between-language competitor, the within-language competitor, and the target an average of 600, 889, and 860 ms, respectively, after the onset of the target word. An example of the eye movements in a simultaneous between- and within-language competition trial is busi ( beaded shown in Figure 1. When instructed to pick up the target object ), the subject made an eye movement to the between-language competi- necklace tor book at 300 ms after the onset of the target word, followed by an eye move- ment to the within-language competitor buben ( tambourine )a t 633 ms after the onset of the target word, before finally looking at the target object 1267 ms after the onset of the target word and picking it up. Control comparisons. Phonetic mapping from sound to object is not the only factor that drives eye movements. Due to purely visual properties, some objects tend to attract a participant’s attention more than others. When selecting our stimuli, we specifically ensured that our phonetically overlapping competitors
11 Applied Psycholinguistics 24:2 183 Marian & Spivey: Bilingual and monolingual processing Figure 1. A view from the scene camera showing the simultaneous between- and within- busi (beaded language competition condition. The target object in the upper left corner is necklace), the within-language competitor in the lower left corner is buben (tambourine), the book ( kniga ), and the filler object between-language competitor in the lower right corner is lipstick in the upper right corner is pomada ). The crosshairs indicate the subject’s fixation ( buben on the within-language competitor . were not more visually attractive than their corresponding fillers. To completely rule out the possibility that our results were driven by such physical properties, control comparisons were performed for between-language, within-language, and simultaneous between- and within-language competition. We compared the proportion of eye movements to competitor objects and their corresponding dis- tractor objects during filler instructions to pick up a target that bore no phonetic similarity to either object name. When the between-language competitor and its corresponding distractor both served as fillers, participants looked at the between-language competitor 16% of the time and at its corresponding filler 19% of the time ( p > .1). When the within-language competitor and its corresponding distractor both served as fil- lers, participants looked at the within-language competitor 19% of the time and at its corresponding filler 27% of the time ( p > .1). When both the between- language competitor and the within-language competitor served as fillers, partic- ipants looked relatively equally at the two competitors and their corresponding fillers. They looked at the between-language competitor 17% of the time and its corresponding filler 19% of the time ( p > .1), and they looked at the within- language competitor 17% of the time and its corresponding filler 27% of the time ( p < .1). Analyses by language showed that participants did not look at
12 Applied Psycholinguistics 24:2 184 Marian & Spivey: Bilingual and monolingual processing the corresponding fillers significantly more than at the competitor objects. The relatively higher overall proportion of looks in the control comparisons was due to the fact that, in these instructions, participants did not fixate on the central cross and therefore made more eye movements to objects in the display overall. These results show that, during filler instructions, participants looked relatively equally at the between-language competitors and their corresponding fillers and at the within-language competitors and their corresponding fillers when there was no phonetic overlap with the target. This suggests that physical differences between the overlapping competitor object and its corresponding filler object were not responsible for the observed between- and within-language competi- tion effects. EXPERIMENT 2 Monolingual English speakers were tested under the same conditions and with exactly the same sets of stimuli (from the English session) to allow direct com- parisons between performance of bilingual speakers in Experiment 1 and perfor- mance of monolingual speakers in Experiment 2. The monolingual speakers were tested in English only and their eye movements were recorded under four conditions: (a) a control condition in which there was no phonetic overlap among any items in the display, (b) a between-language competition condition in which Russian lexical items competed for activation with an English target, (c) a within-language competition condition in which English lexical items com- peted for activation with an English target, and (d) a simultaneous competition condition in which both English and Russian lexical items competed for activa- tion with an English target. The hypotheses were that monolingual English speak- ers will show significant competition from English competitors in both the within- language competition condition and the simultaneous competition condition, but they will show no competition from Russian items in either the between-language competition condition or the simultaneous competition condition. Methods Participants. Twelve English monolinguals participated in the study (5 males, 7 females). All were students at Cornell University; their mean age was 19.6 years ( SD = 0.99). None of them knew Russian or had studied Russian at any time in their lives. Informed consent was obtained and participants’ rights were protected. All participants were paid for their participation. At the end of the experiment, participants were asked whether they had noticed anything unusual and what they thought the objectives of the study were. Normally in this task, any participant who mentions noticing similar sounding object names is ex- cluded from data analysis. In this experiment, none of the participants reported noticing any similarities. Apparatus and stimuli. The same eye-tracking equipment and the same sets of stimuli were used with monolinguals tested in Experiment 2 as with the bilin- guals tested in Experiment 1. Because the English set that included card as a
13 Applied Psycholinguistics 24:2 185 Marian & Spivey: Bilingual and monolingual processing target was dropped from the bilingual analyses, this set was not used with the monolinguals. Design and procedure. The design of Experiment 2 followed that of Experi- ment 1 except that all trials were presented in English only. All participants were tested with 36 English trials, of which 9 were control trials, 9 were be- tween-language competition trials, 9 were within-language competition trials, and 9 were simultaneous between- and within-language competition trials. Analyses. The proportion of eye movements to competitor items was com- pared to the proportion of eye movements to a noncompeting filler item in the same position. As in the bilingual experiment, the between-language competitor and the within-language competitor were necessarily in different locations in the display (and were present simultaneously in one of the conditions) and had differ- ent objects as their corresponding control fillers. Therefore, we conducted separate repeated-measures ANOVAs for eye movements to the between-language compet- itor location and for eye movements to the within-language competitor location. These analyses were performed across all four conditions: the independent vari- ables were presence or absence of the between-language competitor and presence or absence of the within-language competitor, and the dependent variable was proportion of eye movements made to the relevant location in the display. Results Analyses of variance were performed by subjects and by items on proportion of looks to competitor items and to control fillers across all four conditions. Monolingual English speakers looked at the within-language English competitor in 20% of all trials and to nonoverlapping control fillers on 8% of all trials. This difference was significant both by subjects and by items, F 1 (1, 11) = 12.21, p < .01; F 2 (1, 8) = 10.45, p < .05. There was no effect of presence or absence of the between-language competitor and no interaction among variables, indicating that monolinguals made eye movements to the English competitor in the within-language competition condition as well as in the simultaneous competition condition. Similar analyses of variance for looks to the Russian competitor revealed no significant effects. Monolingual English speakers looked at the between-lan- guage Russian competitor in 5% of all trials and to the nonoverlapping control filler in 8% of all trials, F 1 (1, 11) = 3.08, p > .1; F 2 (1, 8) = 1.22, p > .1, with no interaction between variables. These results suggest that English monolin- guals treated the Russian competitor as they would any other nonoverlapping control filler. DISCUSSION Additional analyses were performed across both experiments, treating linguistic status (monolinguals vs. bilinguals) as a between-subject variable to compare performance of both bilinguals and monolinguals on the English part of the
14 Applied Psycholinguistics 24:2 186 Marian & Spivey: Bilingual and monolingual processing experiment. A 2 × 2 ANOVA (language, presence/absence of the between- × 2 language competitor, and presence/absence of the within-language competitor) revealed a main effect of within-language competitor presence. Across both ex- periments, participants looked at the English competitor in 20% of all trials and at the nonoverlapping control filler in 8% of all trials. This difference was sig- 1 (1, 25) = 15.575, nificant both by subjects and by items, < .01; F 2 (1, 8) F p 20.38, = < .01. No interaction between variables was observed. These findings p suggest that both bilinguals and monolinguals experienced competition from items overlapping with the target item in English. For competition from Russian, across both experiments, participants looked at the Russian competitor in 11% of all trials and at the nonoverlapping control F 1 (1, 25) = 7.148, p < filler in 7% of all trials, F 2 (1, 8) = 1.90, p > .1. .05; A significant interaction with linguistic status (monolingual vs. bilingual) was 5.831, observed, (1, 25) = 14.57, p < .01; F 2 (1, 8) = 1 p < .05. This suggests F that the difference in the proportion of eye movements to the Russian competitor and the control filler was an artifact of the interaction and was due to perfor- mance of bilingual participants only. Namely, during the English session, bilin- gual Russian–English speakers looked at the Russian competitor objects in 17% of all trials and at the nonoverlapping control filler in the same location on 6% of all trials, whereas monolingual English speakers looked at the Russian competitor objects in 4% of all trials and at the nonoverlapping control filler in the same location on 7% of all trials. This pattern of eye movements suggests that Russian–English bilinguals made eye movements to the Russian competitor to the same extent as they did to the English competitor. However, English monolingual speakers, not knowing Russian, treated the Russian competitor as any other filler in the display and did not make more eye movements to items overlapping phonetically in Russian. The present study reinforces the hypothesis of parallel activation of two lan- guages during bilingual language processing. The main effect of between-lan- guage competition across languages replicates the basic finding of Spivey and Marian (1999). However, they found that competition was stronger from the second language into the first, while in the present study significant between- language competition was observed from both languages and into both lan- guages. These differences in results suggest that factors such as stimulus selec- tion and the overall degree of activation of a language influence the extent to which languages compete during processing. Ensuring equal phonetic overlap between competitor and target items in both languages, controlling for word frequencies, and boosting activation of the language not used in daily environ- ments resulted in a study that was cleaner and better balanced across languages. These differences suggest that subtle experimental manipulations may alter a bilingual’s performance and influence the pattern of language processing. The finding that the activation of a language and the degree to which it shows inter- ference may be malleable by experimental manipulation is in itself an interesting idea, one that further calls attention to methodological designs of bilingual ex- periments. Moreover, in the future, a careful study in which word frequency and phonetic overlap are independent variables may prove insightful. For example, if competition is tested for more than one item within the same language, word
15 Applied Psycholinguistics 24:2 187 Marian & Spivey: Bilingual and monolingual processing frequency and amount of phonetic overlap are likely to be important determi- nants of strength of competition from each item. Our study was the first to consider eye-tracking evidence for within-language competition in bilinguals. The results suggest that, similar to monolingual En- glish speakers, bilingual Russian–English speakers also encounter within-lan- guage lexical competition. These findings provide further support for the robust- ness of the within-language competition phenomenon. An investigation of simultaneous within- and between-language competition during bilingual spo- ken language processing suggests that competition in bilinguals takes place from both the first and second languages at the same time. The results of the bilingual experiment permit us to conclude that during spoken language processing, the interplay between visual and auditory information processing can cause bilin- gual listeners to encounter competition from items between their two languages as well as within their two languages. This conclusion is further strengthened by the pattern of performance ob- served in the monolingual English speakers tested in Experiment 2 and by the direct comparison between monolingual performance and bilingual performance on the English part of Experiment 1. The fact that monolinguals experience within-language competition is not a new finding in itself (Tanenhaus et al., 1995). What is interesting about Experiment 2 is that the English monolinguals showed within-language competition that was very similar to the bilinguals tested in Experiment 1, but they showed a drastic difference for between-lan- guage competition. These similarities in performance for within-language com- petition and differences in performance for between-language competition are interpreted as supporting parallel activation at the phonological level of the two languages in bilinguals. They suggest that the between-language competition effect is genuine and not an artifact of biased stimulus selection or design. To conclude, the results of our study provide strong support for parallel spo- ken language processing in bilinguals. We observed robust between-language and within-language competition effects in both languages. Together with other findings of parallel processing in bilinguals, these results suggest that bilingual listeners simultaneously accumulate phonemic input into both of their lexicons (which presumably cascades to higher levels of representation) as a spoken word unfolds in real time, even when in a monolingual situation. In fact, given the accumulating evidence for cross-linguistic competition (Bijeljac–Babic et al., 1997; Spivey & Marian, 1999; van Heuven et al., 1998), the assumption of two separate lexical modules for word recognition may be challenged. Moreover, explicit tests of competing theories will eventually require computational mod- els of bilingual language processing (e.g., Dijkstra, van Heuven, & Grainger, 1998) with architectures that can accommodate this apparent interaction between the two languages throughout the entire processing stream.
16 1.447 0.06 6.9465 3.59 3.1838 1.74 Word 19.78 27.78 11.2301 13.1404 35.95 Frequency 86.14/54.13 Noncompeting Filler Filler Moose Shoelace Spoon Microphone Screwdriver Lightbulb Toothbrush Key chain Lipstick Pencil 3.76 9.6093 Word 59.07 61.92 96.73 35.5847 11.29 65.41 132.735 Frequency 54.5/77.68 6.95/51.06 216.61/5.96 Filler Razor Tea bag Lizard Sponge Train Plate Ruler Hair clip Perfume Eye 4.978 9.899 5.847 7.294 5.036 67.35 14.819 41.158 12.098 Word 116.05 270.97 301.418 Frequency Within-Language Competitor English Gum Plum Car Bark Chess set Shovel Spear Book Marbles Lobster 4.63 2.83647 2.60492 Word 40.13 40.17 78.08 17.25036 15.86 18.00289 24.14 15.8032 Frequency 260.03 Filler Fish Knife Napkin Pear Disk Mitten Fork Pin Dental floss Watch 9.841 0.0521 Word 21.84 63.213 17.482 39.884 16.89 12.041 20.724 14.645 24.66 Frequency 19.913/17.771 Between-Language Competitor Engl. Transl. of Russ. Nut Dress Tambourine Potato Competitor Matches Balloon Velvet Nail polish Turtle Stamp 9.088 8.336 3.878 0.81 Word 51.809 73.169 32.272 42.373 16.44 95.224 21.592 30.644 Frequency English Target 7. Plug 8. Gun 9. Card 6. Barbed wire 2. Boot 3. Shark 4. Chair 5. Marker 1. Speaker Target Standard deviation WORD FREQUENCIES FOR TARGET, COMPETITOR, AND FILLER ITEMS USED IN THE ENGLISH PART APPENDIX A AND FOR THEIR TRANSLATION EQUIVALENTS IN RUSSIAN 10. Lock Word frequencies in the English language Mean 188
17 7 2 5 6 13 30 41 75 Word Frequency 23 34 23 NA 5/41 NA NA NA NA NA NA NA 1993 1977 26.67 22.5 6.351 23.88 Noncompeting Filler Filler V,H8" 1J$>"b [email protected]<"*" <[email protected]@> :"*"T T>[email protected] :@08" 6 3 7 4 48 10 NA 128 1093 Word Frequency 31 67 NA NA NA NA NA NA 971 1993 1977 NA/10 224 294.25 169.22 452.26 354.61 Filler <,[email protected],8 H"D,:8" 2"[email protected]:8" [email protected],2* :4>,68" *JN4 <@R":8" $D4H&" bV,D4P" (:"2 R"6>Z6 8 9 33 43 22 11 NA NA 691 Word Frequency 21 17 22 53 20 Within-Language Competitor NA NA NA 231 105.13 150.11 162.89 262.68 465/12 523/11 8" / of Engl. D,24>8" Russ. Trans. [email protected]\, 8>4(" :@B"H" T"N<"HZ [email protected]" T"D484 F:4&" 0,&"H,:\>"b <"T4>" D"8 7 12 31 12 16 16 NA 150 5/32 45.7 Word /34 72 61 Frequency NA NA NA NA NA NA NA NA 1993 1977 Competitor 1993 1977 71 61.26 64.24 NA >4H8" Filler *4F8,H" &"D,08" (DJT" &4:8" 2J$>"b $J:"&8" >@0 DZ$8" F":L,H8" R"FZ 3 1 7 83 24 28 15 71 28 1977 28.889 29.26 Word Frequency 32 31 19 74 23 68 Between-Language Competitor NA NA NA 1993 41.7 23.69 10/3304/20 3/3254/18 >@(H,6 FB4R84 $J$,> T"D48 R,D,B"N" <"D8" B:"H\, $"DN"H ("68" 8"[email protected]" :"8 *:b 4 4 7 98 28 18 31 Word 10 65 22 25 21 Frequency NA NA NA NA 106 123 57.5 45 63.12 46.5 19/19 10/27 English Target [email protected]:@>8" F"[email protected]( "8J:" FHJ: L:@<"FH,D [email protected]:`R"b [email protected][email protected]:@8" F:4&" [email protected]:,H @H8DZH8" 2"<@8 deviation 1. 2. 3. 4. 5. 6. 7. 8. 9. Word frequencies in the Russian language Russ. Trans. of Engl. Target 1993 1977 Russian Standard Mean 10. 189
18 19.78 27.78 0.06 6.9465 1.447 3.1838 3.59 1.74 Word 11.2301 13.1404 35.95 Frequency 86.14/54.13 Noncompeting Filler Filler Moose Lightbulb Microphone Shoelace Screwdriver Toothbrush Pencil 3.76 9.6093 59.07 61.92 Word 35.5848 96.73 11.29 65.41 132.735 Frequency 6.95/51.06 Keychain Filler Ruler Tea bag 54.50/77.68 Spoon Razor Lizard Train Sponge Plate Perfume Eye Hair clip 216.61/5.96 Lipstick 3.763 0.058 1.216 9.899 0.521 17.45 24.09 Word 30.333 16.266 17.482 80.289 14.645 Within-Language Competitor Engl. Trans. of Tie Worm Potato Raincoat Hat Knitting needles Velvet Cheesecloth Tambourine Shovel Russ. Competitor Frequency 4.63 2.83647 2.60492 40.13 78.08 Word 15.86 18.00289 17.25036 24.14 15.8032 260.03 Frequency 4017 Filler Fish Napkin Fork Pear Disk Knife Dental floss Pin Mitten Watch 4.978 0.81 3.878 77.92 51.809 95.224 16.44 12.098 21.592 Word 113.64 270.97 301.418 Frequency Between-Language Competitor English Gun Car Barbed wire Marker Chair Shark Spear Plum Book 9.841 4.863 39.421 39.61 39.884 24.66 20.724 63.213 12.041 50.51 Word 183.329 Frequency 19.913/17.771 Lock Russian Target 2. Necklace 1. Matches 8. Nut 9. Map 3. Ballon 7. Dress 5. Stamp 4. Turtle 6. Jar of Russ. Target Engl. Trans. Standard deviation WORD FREQUENCIES FOR TARGET, COMPETITOR, AND FILLER STIMULUS ITEMS USED IN THE RUSSIAN Mean PART AND FOR THEIR TRANSLATION EQUIVALENTS IN ENGLISH Word frequencies in the English language APPENDIX B 10. Nail polish 190
19 7 2 5 6 13 30 41 75 Word Frequency 23 34 23 6.35 23.88 NA NA 5/41 NA NA NA NA NA NA 1993 1977 26.67 22.5 Noncompeting Filler Filler V,H8" [email protected]<"*" 2J$>"b :" @H&,DH8" $D,:@8 :@F\ T>[email protected] :@08" 8"D">*"T 3 6 7 4 48 10 NA 128 224 1093 Word Frequency 31 67 NA NA NA NA NA NA 108 971 1993 1977 294.25 169.22 452.26 354.61 Filler <,[email protected],8 H"D,:8" 2"[email protected]:8" [email protected],2* :4>,68" <@R":8" *JN4 R"6>Z6 $D4H&" bV,D4P" (:"2 2 3 5 2 77 15 21 13 28 33 Word Within-Language Competitor Frequency 62 13 18 34 68 17 NA NA NA NA 35.33 19.9 24.15 22.90 FB4PZ $J$,> T"B8" R,D&b8 $"DN"H (":FHJ8 B:"V <"D:b 8"[email protected]" :@B"H" 7 7 31 12 16 16 NA 150 45.7 Word Frequency 34 5/32 72 61 NA 181 NA NA NA NA NA NA 1993 1977 Russian 1993 1977 71 61.26 64.24 >4H8" Filler *4F8,H" &"D,08" (DJT" &4:8" 2J$>"b DZ$8" >@0 F":L,H8" R"FZ $J:"&8" 9 4 98 28 11 31 NA 691 Word Frequency 21 65 22 21 NA NA NA 231 97.22 131.18 19/9 10/27 1993 1977 154.50 239.63 465/12 523/11 Between-Language Competitor 8" / of Engl. [email protected][email protected]:@8" Competitor Russ. Trans. [email protected]\, "8J:" F"[email protected]( L:@<"FH,D FHJ: F:4&" [email protected]:`R"b [email protected]:,H <"T4>" 2"<@8 6 1 7 83 24 28 71 62 154 1977 48.44 49.77 Word Frequency 32 31 19 74 47 23 83 NA NA 1993 44.14 25.20 10/3304/20 3/3254/18 Russian Target FB4R84 T"D48 $JFZ <"D8" R,D,B"N" B:"H\, $">8" ("68" :"8 *:b 8"DH" >@(H,6 deviation 1. 3. 2. 5. 4. 7. 6. 8. 9. Target Word frequencies in the Russian language 10. Standard Mean 191
20 Applied Psycholinguistics 24:2 192 Marian & Spivey: Bilingual and monolingual processing ACKNOWLEDGMENTS The authors wish to thank the editor and anonymous reviewers for helpful comments and suggestions on an earlier version of this manuscript. We particularly thank Jonathan Grainger for suggesting the monolingual experiment, Catherine Snow for suggesting changes in data analyses, and Ton Dijkstra and Judy Kroll for insightful discussions of this work. We also wish to acknowledge Eugene Shildkrot, Marina Basina, and Olga Kats for their help in constructing stimuli, data collection, and coding and Margarita Kaushanskaya for help with the phonetic counts and word frequencies. REFERENCES Allopenna, P., Magnuson, J., & Tanenhaus, M. (1998). Tracking the time course of spoken word recognition using eye movements: Evidence for continuous mapping models. Journal of Memory and Language, 38, 419–439. Bijeljac–Babic, R., Biardeau, A., & Grainger, J. (1997). Masked orthographic priming in bilingual word recognition. Memory and Cognition, 25, 447–457. Brysbaert, M., Van Dyck, G., & Van de Poel, M. (1999). Visual word recognition in bilinguals: Evidence from masked phonological priming. Journal of Experimental Psychology: Human 137–148. Perception and Performance, 25, Chen, H. C., & Ho, C. (1986). Development of Stroop interference in Chinese–English bilinguals. Journal of Experimental Psychology: Learning, Memory, and Cognition, 12, 397–401. Colome, A. (2001). Lexical activation in bilinguals’ speech production: Language-specific or lan- guage-independent? 721–736. Journal of Memory and Language, 45, DeGroot, A., & Nas, G. (1991). Lexical representation of cognates and noncognates in compound bilinguals. Journal of Memory and Language, 30, 90–123. de Groot, A. M. B., Delmaar, P., & Lupker, S. J. (2000). The processing of interlexical homographs in translation recognition and lexical decision: Support for nonselective access to bilingual memory. , 397–428. Quarterly Journal of Experimental Psychology, 53A Dijkstra, A., Timmermans, M., & Schriefers, H. (1997). Cross-language effects on bilingual homo- graph recognition .U npublished manuscript. Dijkstra, A., van Heuven, W. J. B., & Grainger, J. (1998). Simulating cross-language competition with the bilingual interactive activation model. Psychologica Belgica, 38, 177–196. Dijkstra, A., van Jaarsveld, H., & ten Brinke, S. (1998). Interlingual homograph recognition: Effects of task demands and language intermixing. 51–66. Bilingualism, 1, Dijkstra, T., Grainger, J., & van Heuven, W. J. B. (1999). Recognition of cognates and interlingual homographs: The neglected role of phonology. 496– Journal of Memory and Language, 41, 518. Dijkstra, T., Grainger, J., & van Heuven, W. J. B. (2002). Recognition of cognates and interlingual .M anuscript submitted for publication. homographs Doctor, E. A., & Klein, D. (1992). Phonological processing in bilingual word recognition. In R. J. Harris (Ed.), Cognitive processing in bilinguals (pp. 237–252). Amsterdam: Elsevier. Elman, J. L., & McClelland, J. L. (1986). Exploiting lawful variability in the speech wave. In J. S. Perkell & D. H. Klatt (Eds.), Invariance and variability in speech processes. Hillsdale, NJ: Erlbaum. Gerard, L. D., & Scarborough, D. L. (1989). Language-specific lexical access of homographs by bilin- guals. 30 5– 313. Journal of Experimental Psychology: Learning, Memory, and Cognition, 15, Grainger, J. (1993). Visual word recognition in bilinguals. In R. Schreuder & B. Weltens (Eds.), The bilingual lexicon (pp. 11–26). Amsterdam: John Benjamins. Grainger, J., & Dijkstra, A. (1992). On the representation and use of language information in bilin- guals. In R. J. Harris (Ed.), Cognitive processing in bilinguals (pp. 207–220). Amsterdam: Elsevier. Kroll, J., & Stewart, E. (1994). Category interference in translation and picture naming: Evidence for asymmetric connections between bilingual memory representations. Journal of Memory and Language, 33, 149–174. Lenngren, L. (Ed.). (1993). Chastotnyi slovari sovremennogo Russkogo yazyka [Frequency diction- ary of modern Russian language]. Uppsala, Sweden: Acta Universitatis Upsaliensis.
21 Applied Psycholinguistics 24:2 193 Marian & Spivey: Bilingual and monolingual processing Li, P. (1996). Spoken word recognition of code-switched words by Chinese–English bilinguals. Journal of Memory and Language, 35, 757–774. MacNamara, J., & Kushnir, S. (1971). Linguistic independence of bilinguals: The input switch. 480–487. Journal of Verbal Learning and Verbal Behavior, 10, Marian, V. (2000). Bilingual language processing: Evidence from eye-tracking and functional neu- roimaging (Doctoral dissertation, Cornell University, 2000). Dissertation Abstracts Interna- tional, 61. Cognitive Psychology, McClelland, J., & Elman, J. (1986). The TRACE model of speech perception. 1– 18, 86. Nas, G. (1983). Visual word recognition in bilinguals: Evidence for a cooperation between visual and sound based codes during access to a common lexical store. Journal of Verbal Learning and Verbal Behavior, 22, 526–534. Preston, M., & Lambert, W. (1969). Interlingual interference in a bilingual version of the Stroop Color–Word Task. Journal of Verbal Learning and Verbal Behavior, 8, 295–301. Romaine, S. (1995). Bilingualism. Oxford: Blackwell. Soares, C., & Grosjean, F. (1984). Bilinguals in a monolingual and a bilingual speech mode: The effect on lexical access. 380–386. Memory and Cognition, 12, Spivey, M., & Marian, V. (1999). Crosstalk between native and second languages: Partial activation of an irrelevant lexicon. Psychological Science, 10, 281–284. Tanenhaus, M., Spivey–Knowlton, M., Eberhard, K., & Sedivy, J. (1995). Integration of visual and linguistic information during spoken language comprehension. Science, 268, 1632–1634. van Heuven, W. J. B., Dijkstra, T., & Grainger, J. (1998). Orthographic neighborhood effects in bilingual word recognition. 458–483. Journal of Memory and Language, 39, Zasorina, L. N. (Ed.). (1977). Chastotnyi slovari Russkogo yazyka [Frequency dictionary of Russian language]. Moscow: Russkii Yazyk. Zeno, S., Ivens, S., Millard, R., & Duvvuri, R. (1995). The educator’s word frequency guide .B rew- ster, NY: Touchstone Applied Science Associates.