Bricks or Mortar: Which Parts of the Input Does a Second Language Listener Rely on?

Bricks or Mortar: Which Parts of the Input Does a Second Language Listener Rely on? JOHN FIELD University of Reading Reading, England There is considerable evidence from psycholinguistics that first language listeners handle function words differently from content words This makes intuitive sense because content words require the listener to access a lexical meaning representation whereas function words not A separate channel of processing for functors would enable them to be detected faster The question is of importance to our understanding of second language (L2) listening Because what is extracted from the input by L2 listeners is generally less than complete, it is useful for the instructor to know which parts of the signal they are likely to recognize, and which parts are likely to be lost to them On the one hand, L2 listeners might rely heavily on function words because high frequency renders them familiar On the other, they might have difficulty identifying function words confidently within a piece of connected speech because functors in English are usually brief and of low perceptual prominence The current study investigated intake by intermediatelevel L2 listeners to establish whether function or content words are processed more accurately and reported more frequently It found that the recognition of functors fell significantly behind that of lexical words The finding was remarkably robust across first languages and across levels of proficiency, suggesting that it may reflect the way in which L2 listeners choose to distribute their attention rammarians have long found it useful to identify two categories of lexical unit The distinction is often expressed in terms of a closed class (of prepositions, determiners, auxiliary verbs, etc.) to which new items are very rarely added, and an open class (consisting mainly of nouns, verbs, adjectives, and adverbs of manner) which is constantly being expanded as new items are coined (Quirk, Greenbaum, Leech, & Svartvik, 1985, pp 67–68) A more traditional way of defining the categories is by distinguishing items which fulfil a largely syntactic function (function words) from those which bear lexical meaning (content words) This dis- G TESOL QUARTERLY Vol 42, No 3, September 2008 411 tinction gives rise to grey areas: For example, the preposition in clearly does not have the same level of lexical meaning as the word book, but it still appears in dictionaries and can be demonstrated by a language teacher Nevertheless, it is this semantic distinction which is adopted in the current study, for reasons that will become evident A linguistic concept does not necessarily correspond to a psycholinguistic one Just because a category or structure is recognized in grammar theory, one cannot take it for granted that it has psychological reality, that is, that it plays any part in the way the mind constructs or understands utterances However, within first language (L1) psycholinguistics, a great deal of evidence has accumulated which suggests that function words are processed differently from those that bear lexical meaning Some of the earliest indications came from examples of slips of the tongue, where it was noted that content words are quite often misplaced (rules of word formation → words of rule formation) but that function words tend not to be This evidence suggested to some commentators (e.g., Garrett, 1980) that assembling an utterance demands two distinct processes, with the speaker first constructing a frame in which certain positions are reserved for the mortar of function words and then inserting meaning-bearing bricks in the form of nouns, verbs, and adjectives Further early evidence (for listening, see Swinney, Zurif, & Cutler, 1980) came from patients who had suffered damage to Broca’s area in the brain as a result of a stroke, an accident, or surgery Their vocabulary store seemed to remain relatively intact, but access to grammar (including inflections and function words) was often impaired This led Bradley (1978) to conclude that the two categories are stored separately in the mind and/or accessed in different ways The symptoms associated with Broca’s aphasia might be the result of damage to a part of the brain where function words are grouped or of damage to the route by which the individual retrieves these items when needed It is also suggestive that infants appear to recognize function words quite early in their language development (Shi, Werker, & Cutler, 2006) but not produce many of them until quite late (Radford, 1990), despite their high frequency The delay might be interpreted as an indication that, as speakers, they need to establish a separate retrieval process for this category The long-standing assumption that content and function words are stored and processed differently has not gone unchallenged From tests using spliced sections of speech, Herron and Bates (1997) concluded that, though listeners access function words rapidly, they depend upon context in order to decode them unambiguously Segalowitz and Lane (2000) argued that it is not necessary to assume two stores because the high frequency of function words will always lead to their being recognized more rapidly than most content words However, the separate 412 TESOL QUARTERLY store view has been supported by recent neurological evidence obtained from brain imaging (e.g., Brown, Hagoort, & ter Keurs, 1999; Münte et al., 2001) The areas of the brain that are associated with function word processing appear to be rather different from those associated with accessing the wider lexicon In addition, event-related potentials measuring electrical activity in the brain suggest neurological differences in the way the two categories are processed (Kutas & Van Petten, 1994, pp 125– 127) Using two different routes has practical benefits for a listener (or indeed a reader) In order to identify a function word, a simple pattern match is all that is required; there is no need to gain access to a meaning The process so far as a content word is concerned is considerably more complex The word might possess not one but a range of potential senses, all of which have to be accessed (Swinney, 1979) before one is chosen that accords with the context in which the word appears So a separate route for function words might enable them to be identified more rapidly and less ambivalently than content words If we assume that content and function words are processed separately in this way, then how does a listener, exposed to a group of sounds, manage to determine at the outset which parts of the input are likely to correspond to functors and which to content words? Basing their analysis on English,1 Grosjean and Gee (1987) suggest that listeners exploit the perceptual difference between stressed syllables, which occur almost exclusively in content words (and indeed may even serve as the principal means of identifying those words), and unstressed syllables, which often correspond to monosyllabic, weak quality functors Stressed syllables initiate a lexical search, while unstressed ones lead in the first instance to a simple pattern-matching procedure Grosjean and Gee (1987) assume that there are two separate stores, with one list consisting purely of functors and another of both functors and content words The first is thus available for a rapid identification process, while the second enables factors such as frequency, multiple meaning, and contextual constraints to be brought to bear on any decision The consequence is a much faster identification of function words.2 It should be stressed that any identification is provisional Just because an apparent functor match is achieved, it does not mean that it will be sustained by the subsequent search To give an example, the group The proposal related to English, but many other languages appear to downgrade the prominence of function words in terms of their duration, their loudness, or their vowel quality By including the entire vocabulary, the second list allows for gradations in the extent to which words convey “meaning.” WHICH PARTS OF THE INPUT DOES A SECOND LANGUAGE LISTENER RELY ON? 413 [ə|vз:td]3 might initially be interpreted as HAVE + past participle of verb before being revised when the lexical search reveals that there is no such verb as erted Similarly, a decision on a possible functor match would be sustained in a sequence like [getə|laf] but would be overruled in a sequence such as [|kætəlɒg], once the larger word became available Current models of listening represent lexical retrieval as involving a form of competition with rival candidates receiving activation according to how probable they are as a match, that is, according to how well they fit what has been heard and to their relative frequency Competition can take place across word boundaries, so, halfway through the word catalogue, potential whole word matches such as catalogue and catapult compete with possibilities such as cat + a or cat + of What gives function words a head start, in principle at least, is their high frequency (Segalowitz & Lane, 2000) However, a problem lies in the disproportionately high frequency of many of them,4 which could lead to their dominating content words in which they are embedded (one thinks of the in weather) Bard (1990, p 206) suggests that there is a trade-off between the weak perceptibility of most function words, which depresses their activation, and their high frequency Like Grosjean and Gee, Cutler (1993) suggests that perceptual cues can be used by a listener to initiate separate searches, though she prefers to rely on a distinction between strong stressed syllables with full quality vowels and weak unstressed ones marked by the presence of schwa She points out that the structure of the English lexicon means that any first-pass association of content words with the former and function words with the latter has a high chance of success Of the strong syllables in a corpus examined by Cutler & Carter (1987), 86% occurred in open class words and only 14% in closed-class words The pattern was reversed for weak syllables, with 72% in closed-class words and 28% in open-class words Stankler (reported in Cutler, 1993) supported these statistics with psychological evidence of a link between weak quality and membership of the functor class He trained English speakers in an artificial language which observed the same prosodic distinctions as in English, with weak quality items serving grammatical functions and strong quality items carrying lexical content The results were compared with those from participants who learned the same artificial language but with the content–functor distinction random, reversed (strong syllables marking 414 The phonemic transcription in this article follows the conventions of the International Phonetic Alphabet The examples given are based upon standard southern British English Of the 100 most frequent items in the spoken British National Corpus (Leech, Rayson, & Wilson, 2001, p 144), only around 16 are clearly identifiable as open-class words The word the leads the field, with a frequency of 39,605 per million words, with its nearest rival I at 29,448 By comparison, the first content word know has a frequency of 5,550 per million TESOL QUARTERLY function words), or null (all syllables strong) Stankler reported a significant advantage in favour of the version that followed the same distinction as English The conclusion he drew is that English listeners know and exploit the connection between vowel quality and membership of the two classes A rather different slant on how listeners handle input comes from studies of verbatim recall We know that, after a lapse of time, content words are remembered more accurately than functors The most obvious explanation is that functors decay from memory more quickly once the utterance has been turned into an abstract idea because they are not central to the final meaning But it might also be that any sections of the input that potentially correspond to function words are awarded less attention by the listener at the time they are being heard They are processed more shallowly because they are perceptually weaker and less easy to decode with confidence This explanation would correspond to evidence from reading L1 readers process function words more cursorily (Haberlandt & Graesser, 1989) and skip them more often (Carpenter & Just, 1983) The findings may be partly connected to the shortness of most functors, but readers have also shown themselves less accurate in crossing out a target letter when it occurs in a function word than when it occurs in a content one (Rosenberg, Zurif, Brownell, Garrett, & Bradley, 1985) RELEVANCE TO SECOND LANGUAGE (L2) LISTENING We must now relate these issues to L2 listeners Many other languages besides English distinguish content and function words by means of their relative prominence, so learners will often be familiar with the principle from their L1 But they still need to recognise the specific cues that mark the difference in English, and to develop an association between cue and word category Lexically stressed syllables in English are marked in several ways: by greater duration, loudness, and pitch, and by possessing full quality vowels (Laver, 1994, pp 512–514) Most learners succeed in distinguishing stressed from unstressed syllables quite reliably at an early stage However, the ability to make a content/functor attribution may depend upon the overall rhythm of English—perhaps upon the brevity of unstressed syllables compared to stressed ones Eastman (1993) produces evidence that learners face an important obstacle in distinguishing content words and functors when their L1 does not resemble English rhythmically He suggests that speakers of what are traditionally called syllable-timed languages are at a disadvantage compared with those who speak stressed-timed languages WHICH PARTS OF THE INPUT DOES A SECOND LANGUAGE LISTENER RELY ON? 415 Clearly the ability to identify content words confidently is often compromised by the learner’s limited vocabulary A great deal depends on how accurately the word in question is represented in the listener’s mind and how many possible variants of it the listener is able to recognize (Field, 2008b, see chapter 9) L2 listeners constantly need to make allowance for the fact that a given string of sounds may not correspond to any of the lexical items that they currently know But it is functors that pose the most intriguing questions On the one hand, they are highly frequent They are also limited in English to around 300 single or multiword items We can assume that the intermediate-level listener has encountered most of them many times over and has been able to build up a repertoire of the possible variations to which they may be subject On the other hand, the information in the input that signals the possible presence of a function word is usually (because of its unstressed status) very brief and of low perceptibility It is thus potentially unreliable Discussing L1 listeners, Shillcock and Bard (1993) point out that the high frequency of function words is counterbalanced by the fact that the uncertain evidence provided by the input may lead us to form matches with a large number of possible words: As closed-class words are often pronounced as weak syllables, they are likely to have short, centralized vowels or no vowel at all as well as reduced, imperfectly articulated consonants The set of lexical competitors activated by such poor input could be quite large and yet, because the acoustic evidence is so poor, have no clear front runners (p 182) If this uncertainty obtains with an L1 listener, how much more pronounced it must be with an L2 listener—especially one whose phoneme values in English are shaky There is thus a conflict between high frequency/familiarity and low perceptual evidence It is of interest to see which of the factors prevails for most L2 listeners The relevance is that a great deal of L2 listening is highly strategic in nature In general, listeners succeed in decoding far less of the input than is generally assumed (Field, 2008b, see chapter 15) They are thus quite heavily dependent on compensatory techniques to supply the words or the concepts that they have not succeeded in identifying We know very little about the perceptual information that the inexperienced L2 listener draws upon to support these strategic guesses We know even less about which items within that information are likely to be reliable and which items listeners should be cautious about trusting Hence the question addressed in this study: Do function words or content words feature more reliably in the bottom-up data that becomes available to the listener? Do listeners structure their interpretation of a partially understood piece of spoken input around familiar functors? 416 TESOL QUARTERLY They might so, not as part of a “mortar-first” approach to syntactic processing but simply by virtue of the fact that they know and recognize function words and rate them as dependable This would assist them, as Cutler (1993) suggests, to work out where content words begin and end Just such a procedure has been hypothesized in the early stages of L1 acquisition (Christophe, Guasti, Nespor, Dupoux, & van Ooyen, 1997) Alternatively, listeners adopt a decoding strategy that is primarily semantic, and rely more heavily on perceptually reliable information in the form of lexically stressed syllables that provide cues to meaningbearing words? Intuitively, this seems to be the more likely—because more informative—line of attack But it is a line of attack fraught with dangers Listeners may need to (a) allow for the fact that a targeted content word is not in their vocabulary at all, (b) access not one but a range of possible senses, and (c) relate those senses to the context in which the word appears In relation to (c), it is important to bear in mind that the context in question might be incomplete and fragmented because not everything in the utterance has been successfully decoded DECODING BY L2 LISTENERS Relatively few studies have investigated how much of a piece of natural speech an L2 listener succeeds in matching to words The reasons have been partly practical: Multiple variables need to be considered, among them the sample of speech that is chosen, the familiarity of the speaker’s voice and accent, the listening experience of the participants, and so on The reasons have also been historical In the 1980s and early 1990s, there was a received idea among TESOL practitioners and researchers that the listener’s ability to map from sounds to words was not of primary importance because any lapses at this lower level could be compensated for by the use of contextual information It is only relatively recently (Field, 2008a; Lynch, 2006; Vandergrift, 2004) that thinking has moved on and there has been renewed interest in perceptual processing Even so, much of the focus of attention has been on how the phonology of L1 constrains the perception of L2 at phoneme level (Strange, 1995) One way of investigating the decoding of natural speech is to play a recording to a population of L2 listeners and to ask them to transcribe what they understand The resulting data can be subjected to error analysis and thus provides input to remedial classroom practice It also indicates what linguistic information is available to listeners A number of research studies have adopted a transcription method, often with a primary interest in vocabulary (i.e., content word) recognition Fishman (1980) concluded from responses obtained in a dictation exercise that WHICH PARTS OF THE INPUT DOES A SECOND LANGUAGE LISTENER RELY ON? 417 the errors of L2 listeners were not dissimilar to those of L1 listeners Voss (1984) examined the effects of hesitation and accent on accuracy and on the types of error made There were lexis versus syntax studies by Conrad (1983) and Kelly (1991), though the criteria they used for the subsequent classification of errors are open to challenge Mack (1988) exposed learners to pieces of computer-generated speech and to anomalous utterances such as A painted shoulder thawed the misty sill More recently, Bonk (quoted in Pemberton, 2004) traced the relationship between comprehension and the accuracy with which content words are transcribed, and Pemberton (2004) investigated the decoding of 27 Hong Kong listeners, using a carefully designed taxonomy of error types A brief look at the results of Voss’s first task indicates that, of 193 errors by his 22 high-level German learners, 109 involved function words However, the figures are not as clearcut as they might appear: Some words were wrongly reported by nearly all respondents, others by only one or two The data does not indicate the relationship between erroneous items and those correctly matched or the relative prevalence of functors and content words in the text Pemberton’s detailed analysis included consideration of the effects of word frequency and word category on recognition He reported (pp 41–42) an unexpectedly low recognition rate for high-frequency words, all of them known to the participants He also reported very little difference between recognition rates for function and content words (respectively, 74% and 79%) An important aspect of the transcription method adopted by both Voss and Pemberton is that it permitted participants to rewind as often as they wished in order to check their answers The opportunity for recursion means that the levels of accuracy and the types of error must differ from those that would obtain in real life One might postulate that errors in situations where only one hearing is possible would be higher than those recorded by Pemberton and that the ratio between accurately reported content and functor words might well be different RESEARCH DESIGN Method: Paused Transcription The method used in this study therefore required listeners to report back immediately after hearing the target sections of speech and without the possibility of rewinding Short sections of only four or five words were targeted, with a view to limiting possible memory effects Small-scale pieces of transcription run the risk of directing attention to decoding at word level, thus eliciting a set of processes which not resemble those 418 TESOL QUARTERLY of a normal listening encounter A paused transcription method was therefore adopted In this paradigm, participants are asked to listen to an authentic piece of connected speech Pauses are inserted into the recording at irregular intervals; and, whenever a pause occurs, participants are asked to transcribe the last few words The rationale is that, for most of the recording, participants are listening as they would in real life, for larger-scale meaning When the pause occurs, the most recently heard words remain available for report: There is psycholinguistic evidence (Jarvella, 1971) that we briefly retain a verbatim record of the words we hear until the onset of the following clause Material The recording used was from a set of L2 listening comprehension materials (Underwood, 1975) It consisted of an informal interview with the manager of a cinema, in which he discusses changes in the way cinema is perceived The text was judged to be culturally neutral and within the world experience of all the participants Most or all of the words in it were judged to be within their vocabulary range The sections of the text chosen for transcription combined content and function words and are shown in Table A minimum of 10 seconds of recording separated each section to ensure that higher-level processing took place Participants Nonnative listeners (NNL) were drawn from mixed-nationality classes at an English language school in Cambridge, England The participants were mainly in their late teens or early twenties, and had spent only three weeks in Britain They were in classes graded as intermediate Three members were later dropped from the sample: One had recorded zero responses to all items; one had an entry test score substantially below that of the rest of the group; and for one, there was no record of an entry test TABLE Sections of Recording Targeted for Transcription which changed each week must be an occasion comfortable seats, good sound we’ve lost in the past been brought up on television a higher standard of entertainment have been shown to adults 10 11 12 13 14 we’re providing a service most of their early years make any money on it children having a good time the age of forty-eight they’re staying at home middle-aged type of person WHICH PARTS OF THE INPUT DOES A SECOND LANGUAGE LISTENER RELY ON? 419 score This left 46 participants, who were divided into two groups of 23 on the basis of their scores in the entry test administered by the school At the time of testing, participants were in the third week of their course; the entry test results were thus sufficiently recent to reflect their knowledge of English The first group (NNL1) comprised those who had scores ranging from 30–60; the second (NNL2) comprised those who had scores ranging from 61–80 The participants spoke a range of L1s They included Spanish (n = 12), German (8), Portuguese (5), Korean (4), Italian (4), Japanese (3), Arabic (2), Czech (2), and Mandarin Chinese (2) The group also included one speaker each of French, Russian, Albanian, and Hebrew The issue of possible variation due to native language is explored in due course Control groups of native listener (NL) participants were drawn from Year 10 in two state secondary schools in Cambridge, England They were • • Group NL1: a set of language learners graded as poorly performing (n = 21) Group NL2: a top set of successful language learners (n = 23) In both cases, the language being learnt was French Accurate responses from three members of Group NL1 were found to be well below the mean; these participants were dropped from the sample, on the assumption that they had had writing difficulties with the transcription This left a group of 18 Procedure Participants were tested in groups in their normal classrooms They were told that they would hear a cassette recording of a man’s voice Whenever there was a pause in the recording, they were to write the last four or five words they had heard There was a check to ensure that the instruction had been understood The general specification “four or five” was used so as not to introduce additional cognitive demands by encouraging participants to count the words to be transcribed However, previous experience with the method indicated that, in these circumstances, participants choose to write entire phrases or clauses, which is what happened here The recording (with inserted pauses) was played on high quality audio equipment designed for language learning At each pause, the experimenter called out a number and participants wrote their transcription in the appropriate place on an answer sheet The length of the pauses was designed to ensure that subjects, writing at average speed, could only record a maximum of about six words The aim was to prevent them 420 TESOL QUARTERLY from attempting to recall large sections of the text and, in the process, reducing the accuracy of their responses This expedient also ensured that subjects did not have time to review their answers once they had written them down RESULTS Participants’ handwritten responses were transferred to computer, and were classified word by word according to whether an accurate transcription had been achieved The attitude to spelling was not prescriptive: Orthographic variants were recorded as accurate if they segmented a particular item correctly and approximated phonetically to the target item Lexical words which were accurately identified but wrongly inflected were treated as correct answers Untranscribed words were classified in three ways Where no earlier part of the extract had been transcribed or where the response as a whole consisted of only one word, a blank (–) was recorded Where an earlier part of the text had been transcribed, an omission (0) was recorded A problem was posed by zero responses where the respondent had not transcribed the target string at all or had written words which fell outside the section in question If it could be shown that these responses resulted exclusively from inability to identify words, it would have been legitimate to include them in the figures for errors However, they might equally well have resulted from larger scale failures of processing relating to the whole target section or to the wider context Zero answers were therefore discounted, and accurate transcriptions were calculated as a percentage of those responses where a minimum of one word was written The figures quoted in succeeding sections thus reflect the probability of a response being correct where a response was given Function words in the target items were identified by reference to the specification in Quirk et al (1985, pp 67–72), but the adverb particle up was included and pronoun + verb contractions were treated as single items Though strictly a determiner in its context here, the word most was treated as a content word The compound adjective middle-aged was counted as two items, and the numeral forty-eight was omitted This procedure resulted in a set of 30 content and 29 function words across the 14 items Of the content words, however, three possessed more than two syllables, giving them a distinct advantage in terms of length.5 These longer items (comfortable, television, entertainment) were excluded, leaving 27 content words Since lexical recognition was the goal, the -ing inflection was discounted: The stem in providing is treated as disyllabic and the stems in having and staying as monosyllabic WHICH PARTS OF THE INPUT DOES A SECOND LANGUAGE LISTENER RELY ON? 421 Recognition Across Groups The percentage of words accurately transcribed was calculated for each participant by word type (content vs functor) From these figures, means were calculated for each of the four groups The results are shown in Figure below Results for Group NL2 represent a ceiling effect unsurprising in transcriptions in one’s L1: functor, M = 94.76 (SD = 5.79); content, M = 95.65 (SD = 7.62) Even Group NL1, identified as poorly performing language learners, achieved mean accuracies of above 80%, though there was a striking difference between their ability to transcribe content words and that for functors: functor, M = 84.03 (SD = 13.95); content, M = 91.11 (SD = 6.54) Two main findings emerge Firstly, despite the greater frequency of function words, their recognition rate is considerably lower than that for content words in three of the four groups Secondly, the disparity between the two types of word was relatively consistent across the two NNL groups (17.28 for NNL1, and 19.64 for NNL2) The gap only begins to narrow with the first of the NL groups A series of t tests was used to measure the significance of the differences between function and content word recognition; the results are shown in Table The difference was significant in the cases of both NNL groups; it was borderline in the case of the weak language learner group NL1 To determine the variation between the recognition patterns of the FIGURE Content and functor recognition across four groups 422 TESOL QUARTERLY TABLE Significance of Differences in Content–Functor Recognition Group Nonnative Listener Group Nonnative Listener Group Native Listener Group Native Listener Group Statistic Significance = = = = p < 0.001 p < 0.001 p = 0.056 n.s t(22) t(22) t(17) t(22) 5.77 6.88 2.05 0.64 lower level NNL group (NNL1) and those of the higher level (NNL2), the differential between percentage content-word recognition and percentage function-word recognition was calculated for each participant A t test indicated no significant difference between the groups: t(44) = 0.54, n.s A similar t test comparing differentials between the NL groups also failed to achieve significance: t(39) = 1.80, n.s First Language An uncontrolled variable of this study lay in the mix of L1s within the population sampled The largest native-language groups were therefore investigated independently These were the groups of Spanish speakers (12 participants) and German speakers (8 participants) An important consideration was the fact that one of these languages (German) closely resembles English rhythmically, whereas the other does not.6 Level of English was controlled From the 12 Spanish-speaking participants, eight were chosen whose scores on the school’s entry test were within one point of each of the eight German participants Figure displays the differences in word recognition between the two subgroups Mean percentage recognition by German speakers was higher at 78.75 (11.57) for content words and 61.46 (6.96) for functors This compared with 53.12 (20.34) for content words and 41.14 (17.74) for functors for the Spanish speakers But in both cases, recognition of function words fell well behind that of content words; indeed, it did so to a greater extent in the case of the German speakers A t test on the content– functor differentials across the two language groups indicated no significant difference: t(15) = 0.74, n.s This result suggests that, in identifying weakly stressed items, the Germans did not derive any major advantage from the metrical similarities between L1 and L2 Results were then analysed for three other language groups, each Many commentators regard the traditional stress-timed versus syllable-timed distinction with distrust (see, e.g., Roach, 1982) However, there are marked rhythmic differences between German and Spanish attributable, at least in part, to the prevalence of closed and open syllables and to differences in the relative timing of weak and strong syllables (Delattre, 1965, Dauer, 1983) In both instances, German resembles English more closely WHICH PARTS OF THE INPUT DOES A SECOND LANGUAGE LISTENER RELY ON? 423 FIGURE Mean percentage recognition across two first language groups consisting of four or five individuals at similar levels: These were speakers of Korean, Italian, and Brazilian Portuguese Although the small number of participants means that the results are less reliable, interesting parallels were observed between the content–functor relationship in these languages and that for speakers of Spanish and German (see Figure 3) In most cases, mean percentage recognition of functors was approximately 20% lower for functors than for content words; the exception was Italian, where the gap widened to 29.58% Although these results should be treated with some caution because of the small number of participants in each nationality group, they indicate that greater difficulty in identifying functors in the experimental text was not a function of the participant’s native language The participants featured in this study represent between them two languages sometimes broadly characterized as stress-based (German and Brazilian Portuguese) and three broadly characterized as syllable-based (Spanish, Italian, and Korean) Level of English The evidence reported so far appears to indicate that the accurate identification of function words remains a problem even at higher levels of L2 proficiency The two NNL groups were combined and evidence was sought of a relationship between individual scores on the school’s entry test and accuracy in transcribing the target sections A Pearson product424 TESOL QUARTERLY FIGURE Content–functor accuracy across five first languages moment coefficient suggested a statistically significant relationship between level and content-word recognition (r(44) = 0.64) and a weaker but significant relationship between level and functor recognition (r (44) = 0.58) However, as Figure indicates, large differentials in recognition of the two categories continue in evidence even at the higher levels Figure presents the same data in the form of a scattergram The accuracy with which content words were identified seems to correlate FIGURE Functor–content differential by proficiency level WHICH PARTS OF THE INPUT DOES A SECOND LANGUAGE LISTENER RELY ON? 425 FIGURE Scattergram: Distribution of recognition across proficiency levels quite closely with level as defined by the entry test Clearly, an increase in oral vocabulary might be expected to lead to greater recognition of lexical items However, the ability to identify function words would appear to be subject to considerable individual variation at all levels DISCUSSION The results of this study indicate that English function words are identified significantly less accurately by L2 listeners than are content words The finding was demonstrated across a group of 46 learners of mixed L1s and continued to hold true when participants were controlled for L1 Difficulty in recognising function words also appeared to be independent of level of English Certain learners rated as having a high level of overall proficiency continued to exhibit large differentials between their ability to identify content words and their ability to identify functors This differs markedly from the results recorded by Pemberton (2004), who noted relatively little variation between the levels of content and function words in his data What distinguishes this study from Pemberton’s is that the participants were not able to listen recursively to the recording The present data may thus be more representative of how an L2 listener copes with a piece of natural connected speech in a real-life situation The challenge facing L2 listeners was earlier characterized in terms of a conflict between, on the one hand, the high frequency and limited 426 TESOL QUARTERLY numbers of function words and, on the other, their low perceptual saliency, including their brevity, and (thanks to the ubiquitous schwa) their relatively indeterminate phonetic identity It would appear from the findings that what weighs most with the listener, regardless of level, are perceptual considerations But we need to explore this conclusion in a little more detail One possibility may lie in learners’ unfamiliarity with English phonology: in particular, their difficulty in distinguishing syllables in terms of stress But one would then assume that higher level learners would show a greater facility in identifying function words than lower level ones, given their greater exposure to the target language This was not shown to be the case A second explanation (following Eastman, 1993) might be found in the L1’s rhythmic characteristics: It might be that learners with certain types of L1 are not attuned to discriminating between weak and strong syllables on the grounds, say, of duration or of vowel quality But, again, the evidence did not support such an interpretation Learners with an L1 that can be loosely characterized as stress-based were found to have no advantage over those with other types of L1 prosody (The one counterindication came from the small Italian group) Perception clearly plays a part We cannot ignore the difficulty of making accurate phoneme matches in a foreign language when the syllables are short and uninformative But the results obtained suggest that there is a second contributory factor: namely, the way in which the L2 listener, regardless of level or L1, chooses to distribute his or her attention An earlier section mentioned evidence that L1 users (especially readers) lend a reduced level of attention to function words It is clearly more efficient for an L1 listener to focus primarily on the parts of an utterance which are likely to bear meaning The same logic applies in the case of an L2 listener, but there are additional considerations Listening to a foreign language is, even at higher levels of proficiency, a cognitively demanding activity Working memory is strictly limited in capacity (Gathercole & Baddeley, 1993), and, faced with heavy task demands, language users have to choose where best to direct their attention In these circumstances, the L2 listener opts for the more salient stressed syllables, not simply because content words contribute more to the construction of meaning but also because stressed syllables are more dependable (Bond, 1999) Whereas L1 listeners manage to process functors peripherally because they are so extremely familiar with them (see the results for the NL groups), the L2 listeners not have the capacity to so Their solution is to go for the bricks and let the mortar look after itself The results of this study thus point to a model of L2 listening in which decoding relies primarily upon matching stretches of the speech signal to meaning-bearing items of vocabulary They confirm evidence reWHICH PARTS OF THE INPUT DOES A SECOND LANGUAGE LISTENER RELY ON? 427 ported elsewhere (Field, 2004) of a strategic approach to L2 decoding that appears to be driven by lexical search, even sometimes at the expense of accepting a match that is approximate They also accord closely with VanPatten’s proposal (1994) that the process of making sense of an L2 is principally driven in the early stages of learning by meaning rather than by form APPLICATIONS TO PRACTICE One of the goals of this study was to provide information for the listening instructor If we are to train learners to make sense of partially understood utterances, it is clearly useful to have some notion of the raw material with which the learner is working The evidence suggests that the linguistic foundation on which learners base hypotheses about a speaker’s meaning is likely to consist principally of content words, with the number increasing as knowledge of L2 improves Function words are likely to be missing or only approximately matched, even among some higher level learners These results have implications for the way we instruct learners in listening strategies In the early stages of listening development, learners should be asked to build a general and sometimes approximate meaning representation on the basis of the more prominent content words in the text The current practice of listening out for key words would seem to be justified But as listening competence improves, instructors might move on from meanings to forms: asking learners to attempt to infer the precise words that they had failed to decode (not least among them the absent functors) If, as has been suggested, the inaccurate reporting of functors reflects a bias in the way learners choose to distribute their attention, then we may have to reconcile ourselves to the fact that functors may only ever be decoded in an indeterminate way, even at later levels of L2 proficiency In that case, it makes sense to train learners to infer post hoc what was most likely there A further line of development would be to encourage learners to shift their attention from single, isolated content words to the larger chunks of language within which they regularly occur (Wray, 2002) Thus, one might be satisfied at lower levels with a procedure that identified two occurrences of the word time in quick succession and inferred the likelihood of the sequence from time to time, but at later stages one would expect the entire formula to be produced on a first pass The listeners in this study clearly had their priorities right If one has to choose what to focus on, the ability to identify the meaning-bearing components of an utterance clearly outweighs the benefits of recognis428 TESOL QUARTERLY ing how they are linked syntactically Nevertheless, fine distinctions of meaning can sometimes depend on functor recognition: Compare I’m looking at the photos with I’m looking for the photos It therefore makes sense to include function words in the kind of focused listening practice that some commentators recommend, alongside other phonetically problematic features of the spoken language, such as assimilation, resyllabification, or elision (Field, 2003, 2008b) In this kind of small-scale awarenessraising exercise, a set of (say) 10 sentences containing weak quality realisations of functors might be dictated for transcription The issues that have been raised have implications that extend beyond the teaching of listening A general presumption exists in TESOL that the language taught in the classroom will sooner or later be reinforced by the language encountered outside Exposure to English speakers in person or through film, TV, or podcast affords learners multiple examples of natural language in use The principal requirement for progress, according to many commentators, is that learners should “notice” the form of words that native speakers use or “notice the gap” between their own formulations and the standard one (Schmidt & Frota, 1986) The noticing hypothesis is a persuasive one, but it does not allow fully for possible divergences between what features in the input and what becomes intake to the listener In particular, it does not allow for what has been raised in this study: the possibility that L2 listener attention is directed preferentially toward certain parts of the input and not toward others If we accept this view of events, then it means that many of the less perceptually prominent parts of what learners hear in casual listening may not be present for them or may only be represented in their minds in an indeterminate way In practice, the intake extracted by the learner may be lacking in precisely those elements which might support the development of a complete native-like system of syntax: evidence of inflections and verb forms and of the contribution of function words to larger syntactic patterns This is an issue that clearly merits further consideration ACKNOWLEDGEMENTS I put on record my gratitude to the staff and students at Eurocentre Cambridge and especially to the previous principal, Eryl Griffiths, for assistance with this study I also extend sincere thanks to the staff and pupils at two Cambridge secondary schools, Manor Community College and the Netherhall School, for their support and cooperation THE AUTHOR John Field teaches psycholinguistics and child language development at the University of Reading, England, and cognitive approaches to second language acquisition at WHICH PARTS OF THE INPUT DOES A SECOND LANGUAGE LISTENER RELY ON? 429 Cambridge University, England His interests lie in applying psycholinguistic theory to issues in L2 learning, especially listening He is committed to making psycholinguistics available to a wider audience through his writing and teaching REFERENCES Bard, E (1990) Competition, lateral inhibition and frequency In Altmann, G (Ed.), Cognitive models of speech processing: Psycholinguistic and computational perspectives (pp 185–210) Cambridge, MA: MIT Press Bond, Z (1999) Slips of the ear: Errors in the perception of casual conversation San Diego, CA: Academic Press Bradley, D (1978) Computational distinctions of vocabulary type Unpublished doctoral thesis, Massachusetts Institute of Technology, Cambridge, MA, United States Brown, C M., Hagoort, P., & ter Keurs, M (1999) Electrophysiological signatures of visual lexical processing: Open and closed class words Journal of Cognitive Neuroscience, 11, 261–281 Carpenter, P A., & Just, M A (1983) What your eyes while your mind is reading In K Rayner (Ed.), Eye movements in reading: Perceptual and language processes (pp 275–307) New York: Academic Press Christophe, A., Guasti, Y., Nespor, M., Dupoux, E., & van Ooyen, B (1997) Reflections on phonological bootstrapping: Its role for lexical and syntactic acquisition Language and Cognitive Processes, 12, 585–612 Conrad, L (1983) Semantic versus syntactic cues in listening comprehension Studies in Second Language Acquisition, 7, 59–69 Cutler, A., & Carter, D M (1987) The predominance of strong initial syllables in the English vocabulary Computer Speech & Language, 2, 133–142 Cutler, A (1993) Phonological cues to open- and closed-class words Journal of Psycholinguistic Research, 22, 109–131 Dauer, R M (1983) Stress-timing and syllable-timing reanalyzed Journal of Phonetics, 11, 51–62 Delattre, P (1965) Comparing the phonetic features of English, French, German and Spanish: An interim report Heidelberg, Germany: Julius Groos Eastman, J K (1993) C-words and F-words: The importance of distinguishing content and function in teaching second language listening comprehension System, 21, 495–502 Field, J (2003) Promoting perception: Lexical segmentation in L2 listening ELT Journal, 57, 325–334 Field, J (2004) An insight into listeners’ problems: Too much bottom-up or too much top-down? System, 32, 363–377 Field, J (2008a) Emergent and divergent: A view of second language listening research System, 36, 2–9 Field, J (2008b) Listening in the language classroom Cambridge: Cambridge University Press Fishman, M (1980) We all make the same mistakes: A comparative study of native and non-native errors in taking dictation In J W Oller, Jr., & K Perkins (Eds.), Research in language testing (pp 187–194) Rowley, MA: Newbury House Garrett, M F (1980) Levels of processing in sentence production In B Butterworth (Ed.), Language production: Vol Speech and talk (pp 177–220) London: Academic Press Gathercole, S E., & Baddeley, A (1993) Working memory and language Hove, England: Erlbaum 430 TESOL QUARTERLY Grosjean, F., & Gee, J (1987) Prosodic structure and spoken word recognition Cognition, 25, 135–155 Haberlandt, K., & Graesser, A C (1989) Processing of new arguments at clause boundaries Memory & Cognition, 17, 186–193 Herron, D., & Bates, E (1997) Sentential and acoustic factors in the recognition of open- and closed-class words Journal of Memory and Language, 37, 217–239 Jarvella, R J (1971) Syntactic processing of connected speech Journal of Verbal Learning and Verbal Behavior, 10, 409–416 Kelly, P (1991) Lexical ignorance: The main obstacle to listening comprehension with advanced foreign language learners International Review of Applied Linguistics, 29, 135–149 Kutas, M., & Van Petten, C K (1994) Psycholinguistics electrified: Event-related brain potential investigations In M.-A Gernsbacher (Ed.), Handbook of psycholinguistics (pp 83–143) San Diego, CA: Academic Press Laver, J (1994) Principles of phonetics Cambridge: Cambridge University Press Leech, G., Rayson, P., & Wilson, A (2001) Word frequencies in written and spoken English: Based on the British National Corpus London: Longman Lynch, T (2006) Academic listening: Marrying top and bottom In E Usó-Juan & A Martinez-Flor (Eds.), Current trends in learning and teaching the four skills within a communicative framework (pp 91–110) Amsterdam: Mouton Mack, M (1988) Sentence processing by non-native speakers of English: Evidence from the perception of natural and computer-generated anomalous L2 sentences Journal of Neurolinguistics, 3, 293–316 Münte, T F., Wieringa, B M., Weyerts, H., Szentkuti, A., Matzke, M., & Johannes, S (2001) Differences in brain potentials to open and closed class words: Class and frequency effects Neuropsychologia, 39, 91–102 Pemberton, R (2004) Spoken word recognition in a second language [Research Reports, Vol 6] Hong Kong SAR, China: Hong Kong University of Science and Technology Quirk, R., Greenbaum, S., Leech, G., & Svartvik, J (1985) A comprehensive grammar of the English language London: Longman Radford, A (1990) Syntactic theory and the acquisition of English syntax Oxford: Blackwell Roach, P (1982) On the distinction between “stress-timed” and “syllable-timed” languages In D Crystal (Ed.), Linguistic controversies: Essays in linguistic theory and practice in honour of F R Palmer (pp 73–79) London: Arnold Rosenberg, B., Zurif, E., Brownell, H., Garrett, M., & Bradley, D (1985) Grammatical class effects in relation to normal and aphasic sentence processing Brain and Language, 26, 287–303 Schmidt, R., & Frota, S (1986) Developing basic conversational ability in a second language A case study of an adult learner of Portuguese In R Day (Ed.), Talking to learn: Conversation in second language acquisition (pp 237–326) Rowley, MA: Newbury House Segalowitz, S J., & Lane, K (2000) Lexical access of function versus content words Brain and Language, 75, 376–389 Shi, R., Werker, J F., & Cutler, A (2006) Recognition and representation of function words in English-learning infants Infancy, 10, 187–198 Shillcock, R C., & Bard, E G (1993) Modularity and the processing of closed-class words In G T M Altmann & R C Shillcock (Eds.), Cognitive models of speech processing: The second Sperlonga meeting (pp 163–183) Mahwah, NJ: Erlbaum Strange, W (1995) (Ed.) Speech perception and linguistic experience Baltimore: York Press WHICH PARTS OF THE INPUT DOES A SECOND LANGUAGE LISTENER RELY ON? 431 Swinney, D A (1979) Lexical access during sentence comprehension: (Re)consideration of context effects Journal of Verbal Learning and Verbal Behavior, 18, 645– 659 Swinney, D A., Zurif, E B., & Cutler, A (1980) Effects of sentential stress and word class upon comprehension on Broca’s aphasics Brain and Language, 10, 132–144 Underwood, M (1975) Listen to this! Oxford: Oxford University Press Vandergrift, L (2004) Listening to learn or learning to listen Annual Review of Applied Linguistics, 24, 3–25 VanPatten, B (1994) Evaluating the role of consciousness in second language acquisition: Terms, linguistic features and research methodology AILA Review, 11, 27–36 Voss, B (1984) Slips of the ear Tübingen, Germany: Gunter Narr Wray, A (2002) Formulaic language and the lexicon Cambridge: Cambridge University Press 432 TESOL QUARTERLY ... differently has not gone unchallenged From tests using spliced sections of speech, Herron and Bates (1997) concluded that, though listeners access function words rapidly, they depend upon context in... hypothesized in the early stages of L1 acquisition (Christophe, Guasti, Nespor, Dupoux, & van Ooyen, 1997) Alternatively, listeners adopt a decoding strategy that is primarily semantic, and rely more... 275–307) New York: Academic Press Christophe, A., Guasti, Y., Nespor, M., Dupoux, E., & van Ooyen, B (1997) Reflections on phonological bootstrapping: Its role for lexical and syntactic acquisition