Publications

Abstract

Not only can the pitfalls that Firestone & Scholl (F&S) identify be generalised across multiple studies within the field of visual perception, but also they have general application outside the field wherever perceptual and cognitive processing are compared. We call attention to the widespread susceptibility of research on the perception of speech to versions of the same pitfalls.

Abstract

Speech perception involves prediction, but how is that prediction implemented? In cognitive models prediction has often been taken to imply that there is feedback of activation from lexical to pre-lexical processes as implemented in interactive-activation models (IAMs). We show that simple activation feedback does not actually improve speech recognition. However, other forms of feedback can be beneficial. In particular, feedback can enable the listener to adapt to changing input, and can potentially help the listener to recognise unusual input, or recognise speech in the presence of competing sounds. The common feature of these helpful forms of feedback is that they are all ways of optimising the performance of speech recognition using Bayesian inference. That is, listeners make predictions about speech because speech recognition is optimal in the sense captured in Bayesian models.

Abstract

Listeners learn from their past experience of listening to spoken words, and use this learning to maximise the efficiency of future word recognition. This paper summarises evidence that the facilitatory effects of drawing on past experience are mediated by abstraction, enabling learning to be generalised across new words and new listening situations. Phoneme category retuning, which allows adaptation to speaker-specific articulatory characteristics, is generalised on the basis of relatively brief experience to words previously unheard from that speaker. Abstract knowledge of prosodic regularities is applied to recognition even of novel words for which these regularities were violated. Prosodic word-boundary regularities drive segmentation of speech into words independently of the membership of the lexical candidate set resulting from the segmentation operation. Each of these different cases illustrates how abstraction from past listening experience has contributed to the efficiency of lexical recognition.

Abstract

The phoneme detection task is widely used in spoken-word recognition research. Alphabetically literate participants, however, are more used to explicit representations of letters than of phonemes. The present study explored whether phoneme detection is sensitive to how target phonemes are, or may be, orthographically realized. Listeners detected the target sounds [b, m, t, f, s, k] in word-initial position in sequences of isolated English words. Response times were faster to the targets [b, m, t], which have consistent word-initial spelling, than to the targets [f, s, k], which are inconsistently spelled, but only when spelling was rendered salient by the presence in the experiment of many irregularly spelled filler words. Within the inconsistent targets [f, s, k], there was no significant difference between responses to targets in words with more usual (foam, seed, cattle) versus less usual (phone, cede, kettle) spellings. Phoneme detection is thus not necessarily sensitive to orthographic effects; knowledge of spelling stored in the lexical representations of words does not automatically become available as word candidates are activated. However, salient orthographic manipulations in experimental input can induce such sensitivity. We attribute this to listeners' experience of the value of spelling in everyday situations that encourage phonemic decisions (such as learning new names)

Abstract

Adverse listening conditions always make the perception of speech harder, but their deleterious effect is far greater if the speech we are trying to understand is in a non-native language. An imperfect signal can be coped with by recourse to the extensive knowledge one has of a native language, and imperfect knowledge of a non-native language can still support useful communication when speech signals are high-quality. But the combination of imperfect signal and imperfect knowledge leads rapidly to communication breakdown. This phenomenon is undoubtedly well known to every reader of Speech Communication from personal experience. Many readers will also have a professional interest in explaining, or remedying, the problems it produces. The journal’s readership being a decidedly interdisciplinary one, this interest will involve quite varied scientific approaches, including (but not limited to) modelling the interaction of first and second language vocabularies and phonemic repertoires, developing targeted listening training for language learners, and redesigning the acoustics of classrooms and conference halls. In other words, the phenomenon that this special issue deals with is a well-known one, that raises important scientific and practical questions across a range of speech communication disciplines, and Speech Communication is arguably the ideal vehicle for presentation of such a breadth of approaches in a single volume. The call for papers for this issue elicited a large number of submissions from across the full range of the journal’s interdisciplinary scope, requiring the guest editors to apply very strict criteria to the final selection. Perhaps unique in the history of treatments of this topic is the combination represented by the guest editors for this issue: a phonetician whose primary research interest is in second-language speech (MLGL), an engineer whose primary research field is the acoustics of masking in speech processing (MC), and a psychologist whose primary research topic is the recognition of spoken words (AC). In the opening article of the issue, these three authors together review the existing literature on listening to second-language speech under adverse conditions, bringing together these differing perspectives for the first time in a single contribution. The introductory review is followed by 13 new experimental reports of phonetic, acoustic and psychological studies of the topic. The guest editors thank Speech Communication editor Marc Swerts and the journal’s team at Elsevier, as well as all the reviewers who devoted time and expert efforts to perfecting the contributions to this issue.

Abstract

If listening in adverse conditions is hard, then listening in a foreign language is doubly so: non-native listeners have to cope with both imperfect signals and imperfect knowledge. Comparison of native and non-native listener performance in speech-in-noise tasks helps to clarify the role of prior linguistic experience in speech perception, and, more directly, contributes to an understanding of the problems faced by language learners in everyday listening situations. This article reviews experimental studies on non-native listening in adverse conditions, organised around three principal contributory factors: the task facing listeners, the effect of adverse conditions on speech, and the differences among listener populations. Based on a comprehensive tabulation of key studies, we identify robust findings, research trends and gaps in current knowledge.

Abstract

L2 listening can involve the phantom activation of words which are not actually in the input. All spoken-word recognition involves multiple concurrent activation of word candidates, with selection of the correct words achieved by a process of competition between them. L2 listening involves more such activation than L1 listening, and we report two studies illustrating this. First, in a lexical decision study, L2 listeners accepted (but L1 listeners did not accept) spoken non-words such as groof or flide as real English words. Second, a priming study demonstrated that the same spoken non-words made recognition of the real words groove, flight easier for L2 (but not L1) listeners, suggesting that, for the L2 listeners only, these real words had been activated by the spoken non-word input. We propose that further understanding of the activation and competition process in L2 lexical processing could lead to new understanding of L2 listening difficulty.

Abstract

Speech processing by human listeners derives meaning from acoustic input via intermediate steps involving abstract representations of what has been heard. Recent results from several lines of research are here brought together to shed light on the nature and role of these representations. In spoken-word recognition, representations of phonological form and of conceptual content are dissociable. This follows from the independence of patterns of priming for a word's form and its meaning. The nature of the phonological-form representations is determined not only by acoustic-phonetic input but also by other sources of information, including metalinguistic knowledge. This follows from evidence that listeners can store two forms as different without showing any evidence of being able to detect the difference in question when they listen to speech. The lexical representations are in turn separate from prelexical representations, which are also abstract in nature. This follows from evidence that perceptual learning about speaker-specific phoneme realization, induced on the basis of a few words, generalizes across the whole lexicon to inform the recognition of all words containing the same phoneme. The efficiency of human speech processing has its basis in the rapid execution of operations over abstract representations.

Abstract

Speech recognition in noise is harder in second (L2) than first languages (L1). This could be because noise disrupts speech processing more in L2 than L1, or because L1 listeners recover better though disruption is equivalent. Two similar prior studies produced discrepant results: Equivalent noise effects for L1 and L2 (Dutch) listeners, versus larger effects for L2 (Spanish) than L1. To explain this, the latter experiment was presented to listeners from the former population. Larger noise effects on consonant identification emerged for L2 (Dutch) than L1 listeners, suggesting that task factors rather than L2 population differences underlie the results discrepancy.

Abstract

The acquisition of novel phonetic categories is hypothesized to be affected by the distributional properties of the input, the relation of the new categories to the native phonology, and the availability of supervision (feedback). These factors were examined in four experiments in which listeners were presented with novel categories based on vowels of Dutch. Distribution was varied such that the categorization depended on the single dimension duration, the single dimension frequency, or both dimensions at once. Listeners were clearly sensitive to the distributional information, but unidimensional contrasts proved easier to learn than multidimensional. The native phonology was varied by comparing Spanish versus American English listeners. Spanish listeners found categorization by frequency easier than categorization by duration, but this was not true of American listeners, whose native vowel system makes more use of duration-based distinctions. Finally, feedback was either available or not; this comparison showed supervised learning to be significantly superior to unsupervised learning.

Abstract

To segment continuous speech into its component words, listeners make use of language rhythm; because rhythm differs across languages, so do the segmentation procedures which listeners use. For each of stress-, syllable-and mora-based rhythmic structure, perceptual experiments have led to the discovery of corresponding segmentation procedures. In the case of mora-based rhythm, similar segmentation has been demonstrated in the otherwise unrelated languages Japanese and Telugu; segmentation based on syllable rhythm, however, has been previously demonstrated only for European languages from the Romance family. We here report two target detection experiments in which Korean listeners, presented with speech in Korean and in French, displayed patterns of segmentation like those previously observed in analogous experiments with French listeners. The Korean listeners' accuracy in detecting word-initial target fragments in either language was significantly higher when the fragments corresponded exactly to a syllable in the input than when the fragments were smaller or larger than a syllable. We conclude that Korean and French listeners can call on similar procedures for segmenting speech, and we further propose that perceptual tests of speech segmentation provide a valuable accompaniment to acoustic analyses for establishing languages' rhythmic class membership.

Abstract

Languages differ in the constitution of their phonemic repertoire and in the relative distinctiveness of phonemes within the repertoire. In the present study, we asked whether such differences constrain spoken-word recognition, via two word reconstruction experiments, in which listeners turned non-words into real words by changing single sounds. The experiments were carried out in Dutch (which has a relatively balanced vowel-consonant ratio and many similar vowels) and in Spanish (which has many more consonants than vowels and high distinctiveness among the vowels). Both Dutch and Spanish listeners responded significantly faster and more accurately when required to change vowels as opposed to consonants; when allowed to change any phoneme, they more often altered vowels than consonants. Vowel information thus appears to constrain lexical selection less tightly (allow more potential candidates) than does consonant information, independent of language-specific phoneme repertoire and of relative distinctiveness of vowels.

Abstract

Spraak is continu, er zijn geen betrouwbare signalen waardoor de luisteraar weet waar het ene woord eindigt en het volgende begint. Voor volwassen luisteraars is het segmenteren van gesproken taal in afzonderlijke woorden dus niet onproblematisch, maar voor een kind dat nog geen woordenschat bezit, vormt de continuïteit van spraak een nog grotere uitdaging. Desalniettemin produceren de meeste kinderen hun eerste herkenbare woorden rond het begin van het tweede levensjaar. Aan deze vroege spraakproducties gaat een formidabele perceptuele prestatie vooraf. Tijdens het eerste levensjaar - met name gedurende de tweede helft - ontwikkelt de spraakperceptie zich van een algemeen fonetisch discriminatievermogen tot een selectieve gevoeligheid voor de fonologische contrasten die in de moedertaal voorkomen. Recent onderzoek heeft verder aangetoond dat kinderen, lang voordat ze ook maar een enkel woord kunnen zeggen, in staat zijn woorden die kenmerkend zijn voor hun moedertaal te onderscheiden van woorden die dat niet zijn. Bovendien kunnen ze woorden die eerst in isolatie werden aangeboden herkennen in een continue spraakcontext. Het dagelijkse taalaanbod aan een kind van deze leeftijd maakt het in zekere zin niet gemakkelijk, bijvoorbeeld doordat de meeste woorden niet in isolatie voorkomen. Toch wordt het kind ook wel houvast geboden, onder andere doordat het woordgebruik beperkt is.

Abstract

Dutch-learning and English-learning 9-month-olds were tested, using the Headturn Preference Procedure, for their ability to segment Dutch words with strong/weak stress patterns from fluent Dutch speech. This prosodic pattern is highly typical for words of both languages. The infants were familiarized with pairs of words and then tested on four passages, two that included the familiarized words and two that did not. Both the Dutch- and the English-learning infants gave evidence of segmenting the targets from the passages, to an equivalent degree. Thus, English-learning infants are able to extract words from fluent speech in a language that is phonetically different from English. We discuss the possibility that this cross-language segmentation ability is aided by the similarity of the typical rhythmic structure of Dutch and English words.

Abstract

The central thesis of the target article was that feedback is never necessary in spoken word recognition. The commentaries present no new data and no new theoretical arguments which lead us to revise this position. In this response we begin by clarifying some terminological issues which have lead to a number of significant misunderstandings. We provide some new arguments to support our case that the feedforward model Merge is indeed more parsimonious than the interactive alternatives, and that it provides a more convincing account of the data than alternative models. Finally, we extend the arguments to deal with new issues raised by the commentators such as infant speech perception and neural architecture.

Abstract

Top-down feedback does not benefit speech recognition; on the contrary, it can hinder it. No experimental data imply that feedback loops are required for speech recognition. Feedback is accordingly unnecessary and spoken word recognition is modular. To defend this thesis, we analyse lexical involvement in phonemic decision making. TRACE (McClelland & Elman 1986), a model with feedback from the lexicon to prelexical processes, is unable to account for all the available data on phonemic decision making. The modular Race model (Cutler & Norris 1979) is likewise challenged by some recent results, however. We therefore present a new modular model of phonemic decision making, the Merge model. In Merge, information flows from prelexical processes to the lexicon without feedback. Because phonemic decisions are based on the merging of prelexical and lexical information, Merge correctly predicts lexical involvement in phonemic decisions in both words and nonwords. Computer simulations show how Merge is able to account for the data through a process of competition between lexical hypotheses. We discuss the issue of feedback in other areas of language processing and conclude that modular models are particularly well suited to the problems and constraints of speech recognition.

Abstract

Infants acquire whatever language is spoken in the environment into which they are born. The mental capability of the newborn child is not biased in any way towards the acquisition of one human language rather than another. Because psychologists who attempt to model the process of language comprehension are interested in the structure of the human mind, rather than in the properties of individual languages, strategies which they incorporate in their models are presumed to be universal, not language-specific. In other words, strategies of comprehension are presumed to be characteristic of the human language processing system, rather than, say, the French, English, or Igbo language processing systems. We report here, however, on a comprehension strategy which appears to be used by native speakers of French but not by native speakers of English.

Abstract

Spontaneous self-corrections in speech pose a communication problem; the speaker must make clear to the listener not only that the original Utterance was faulty, but where it was faulty and how the fault is to be corrected. Prosodic marking of corrections - making the prosody of the repair noticeably different from that of the original utterance - offers a resource which the speaker can exploit to provide the listener with such information. A corpus of more than 400 spontaneous speech repairs was analysed, and the prosodic characteristics compared with the syntactic and semantic characteristics of each repair. Prosodic marking showed no relationship at all with the syntactic characteristics of repairs. Instead, marking was associated with certain semantic factors: repairs were marked when the original utterance had been actually erroneous, rather than simply less appropriate than the repair; and repairs tended to be marked more often when the set of items encompassing the error and the repair was small rather than when it was large. These findings lend further weight to the characterization of accent as essentially semantic in function.