Detection of errors during speech production: a review of speech monitoring models

In this paper three theories of speech monitoring are evaluated. The perception-based approach proposes that the same mechanism employed in understanding other-produced language, the speech comprehension system, is also used to monitor one's own speech production. A conceptual, an inner, and an auditory loop convey information to a central, conscious monitor which scrutinizes the adequacy of the ongoing speech flow. In this model, only the end-products in the speech production sequences, the preverbal (propositional) message, the phonetic plan, and the auditory results, are verified. The production-based account assumes multiple local, autonomous monitoring devices, which can look inside formulation components. Moreover, these devices might be tuned to various signals from the actual speech motor execution, e.g. efferent, tactile, and proprioceptive feedback. Finally, node structure theory views error detection as a natural outflow of the activation patterns in the node system for speech production. Errors result in prolonged activation of uncommitted nodes, which in turn may incite error awareness. The approaches differ on the points of consciousness, volition and control, the number of monitoring channels, and their speed, flexibility, and capacity, and whether they can account for concurrent language comprehension disorders. From the empirical evidence presently available, it is argued for a central perception-based monitor, potentially augmented with a few automatic, production-based error detection devices.     

Psychoticism and perceptual processing speed: A test of Robinson's theory

Robinson has argued that bias in the process of knowledge acquisition can give rise to either a differentiated or an integrated learning style. Individuals high on the personality dimension of psychoticism are said to have a differentiated learning style that manifests in superior performance on cognitive tasks that require differentiation of the elements or parts of immediate perceptual experience. To test this theory, 106 undergraduate students aged between 18 and 31 completed the Eysenck Personality Questionnaire—Revised, along with two computer‐based measures of visual processing speed: reaction time (RT) and inspection time (IT). In accord with Robinson's theory, psychoticism scores were found to be inversely related to both mean RT and IT. Moreover, multiple regression analyses indicated that the relationship between psychoticism scores and the two processing speed measures was independent of any age or gender effects. It is recommended that future research examine the relationship between psychoticism, and both poles of Robinson's differentiation versus integration dimension.     

A theory of lexical access in speech production

Preparing words in speech production is normally a fast and accurate process. We generate them two or three per second in fluent conversation; and overtly naming a clear picture of an object can easily be initiated within 600 msec after picture onset. The underlying process, however, is exceedingly complex. The theory reviewed in this target article analyzes this process as staged and feed-forward. After a first stage of conceptual preparation, word generation proceeds through lexical selection, morphological and phonological encoding, phonetic encoding, and articulation itself. In addition, the speaker exerts some degree of output control, by monitoring of self-produced internal and overt speech. The core of the theory, ranging from lexical selection to the initiation of phonetic encoding, is captured in a computational model, called WEAVER++. Both the theory and the computational model have been developed in interaction with reaction time experiments, particularly in picture naming or related word production paradigms, with the aim of accounting for the real-time processing in normal word production. A comprehensive review of theory, model, and experiments is presented. The model can handle some of the main observations in the domain of speech errors (the major empirical domain for most other theories of lexical access), and the theory opens new ways of approaching the cerebral organization of speech production by way of high-temporal-resolution imaging.     

The effects of cognitive activity and perceptual details on speech source monitoring

Our purpose in this study was to investigate the effects of cognitive operations and perceptual details on speech source monitoring. In Phase 1, correctly spelled words and anagrams were presented in Expt 1. Words were read aloud by participants, by a same‐sex voice, or by an opposite‐sex voice. Immediately after Phase 1, in Phase 2, participants were asked whether each word had been read aloud by the participants themselves, by a same‐sex voice, or by an opposite‐sex voice. Source discrimination between own speech and that produced by a same‐sex voice was poorer than between own speech and an opposite‐sex voice. In addition, misattribution of the speech of another to one's self increased as the level of cognitive effort required for the task increased. In Expt 2, misattributions to same‐sex voice were assigned ‘know’ responses more frequently and misattributions to one's self were assigned ‘remember’ responses more frequently. These results suggest that qualitative characteristics such as perceptual detail and cognitive operations are differentially influencing misattributions to the self and those to same‐sex voices.     

Bootstrapping the lexicon: a computational model of infant speech segmentation

Prelinguistic infants must find a way to isolate meaningful chunks from the continuous streams of speech that they hear. BootLex, a new model which uses distributional cues to build a lexicon, demonstrates how much can be accomplished using this single source of information. This conceptually simple probabilistic algorithm achieves significant segmentation results on various kinds of language corpora - English, Japanese, and Spanish; child- and adult-directed speech, and written texts; and several variations in coding structure - and reveals which statistical characteristics of the input have an influence on segmentation performance. BootLex is then compared, quantitatively and qualitatively, with three other groups of computational models of the same infant segmentation process, paying particular attention to functional characteristics of the models and their similarity to human cognition. Commonalities and contrasts among the models are discussed, as well as their implications both for theories of the cognitive problem of segmentation itself, and for the general enterprise of computational cognitive modeling.     

Asynchrony tolerance in the perceptual organization of speech

Researchers have claimed that listeners tolerate large temporal distortion when integrating the spectral components of speech. In some estimates, perceivers resolve linguistic attributes at spectral desynchronies as great as the duration of a syllable. We obtained new measures of perceptual tolerance of auditory asynchrony, using sine-wave synthesis in order to require perceivers to resolve the speech stream dynamically. Listeners transcribed sentences in which the tone analogue of a second formant was desynchronized relative to the remaining tones of a sentence, with desynchrony ranging from a 250-msec lead to a 250-msec lag. Intelligibility declined symmetrically from 72% at synchrony to 7% at 6100 msec. This finding of narrow asynchrony tolerance indicates a time-critical feature of the auditory perceptual organization of speech     

Modulation sensitivity in the perceptual organization of speech

In a spoken utterance, a talker expresses linguistic constituents in serial order. A listener resolves these linguistic properties in the rapidly fading auditory sample. Classic measures agree that auditory integration occurs at a fine temporal grain. In contrast, recent studies have proposed that sensory integration of speech occurs at a coarser grain, approximate to the syllable, on the basis of indirect and relatively insensitive perceptual measures. Evidence from cognitive neuroscience and behavioral primatology has also been adduced to support the claim of sensory integration at the pace of syllables. In the present investigation, we used direct performance measures of integration, applying an acoustic technique to isolate the contribution of short-term acoustic properties to the assay of modulation sensitivity. In corroborating the classic finding of a fine temporal grain of integration, these functional measures can inform theory and speculation in accounts of speech perception.     

Toward a perceptual theory of transparency

Theories of perceptual transparency have typically been developed within the context of a physical model that generates the percept of transparency (F. Metelli's episcotister model, 1974b). Here 2 fundamental questions are investigated: (a) When does the visual system initiate the percept of one surface seen through another? (b) How does it assign surface properties to a transparent layer? Results reveal systematic deviations from the predictions of Metelli's model, both for initiating image decomposition into multiple surfaces and for assigning surface attributes. Specifically, results demonstrate that the visual system uses Michelson contrast as a critical image variable to initiate percepts of transparency and to assign transmittance to transparent surfaces. Findings are discussed in relation to previous theories of transparency, lightness, brightness, and contrast-contrast.     

Evaluating a computational model of perceptual grouping by proximity

A formal approach to the phenomenon of perceptual grouping by proximity was investigated. Grouping judgments of random dot patterns were made by the CODE algorithm (van Oeffelen & Vos, 1982) and several related algorithms, and these judgments were compared with subjects’ grouping judgments for the same stimuli. Each algorithm predicted significantly more subject judgments than would be expected by chance. The more subjects agreed on how a given dot pattern should be grouped, the more successful was the algorithms’ ability to match the judgments for that pattern. CODE predicted significantly fewer subject judgments than did some of the other algorithms, largely because of its overemphasis on the extent of interactivity among dots as they are being grouped.     

Subjective confidence in perceptual judgments: a test of the self-consistency model

Two questions about subjective confidence in perceptual judgments are examined: the bases for these judgments and the reasons for their accuracy. Confidence in perceptual judgments has been claimed to rest on qualitatively different processes than confidence in memory tasks. However, predictions from a self-consistency model (SCM), which had been confirmed for general-information questions (Koriat, 2010) and social attitudes (Koriat & Adiv, 2010), are shown to hold true also for perceptual judgments. In SCM, confidence is modeled by the procedure for assessment of statistical level of confidence: For a 2-alternative, forced-choice item, confidence is based on the consistency with which the choice is favored across a sample of representations of the item, and acts as a monitor of the likelihood that a new sample will yield the same choice. Assuming that these representations are drawn from commonly shared populations of representations associated with each item, predictions regarding the basis of confidence were confirmed by results concerning the functions relating confidence and choice latency to interparticipant consensus and to intraparticipant consistency for majority and minority choices. With regard to the confidence-accuracy (C/A) relationship, the consensuality principle, documented for general-knowledge tasks (Koriat, 2008a), was replicated for perceptual judgments: Confidence correlated with the consensuality of the choice rather than with its correctness, suggesting that the C/A correlation is due to the relationship between confidence and self-consistency and is positive only as long as the correct choices are the consistently made choices. SCM provides a general model for the basis and accuracy of confidence judgments across different domains.     

Enhanced perceptual processing of speech in autism

Theories of autism have proposed that a bias towards low-level perceptual information, or a featural/surface-biased information-processing style, may compromise higher-level language processing in such individuals. Two experiments, utilizing linguistic stimuli with competing low-level/perceptual and high-level/semantic information, tested processing biases in children with autism and matched controls. Whereas children with autism exhibited superior perceptual processing of speech relative to controls, and showed no evidence of either a perceptual or semantic processing bias, controls showed a tendency to process speech semantically. The data provide partial support to the perceptual theories of autism. It is additionally proposed that the pattern of results may reflect different patterns of attentional focusing towards single or multiple stimulus cues in speech between children with autism and controls.     

Reflex theory in a linguistic context: Sergej M. Dobrogaev on the social nature of speech production

The development of reflex theory in its Pavlovian interpretation had significant resonance in a wide range of academic research areas. Its impact on the so-called humanities was, perhaps, no less than the effect it had in medical science. The idea of the conditioned reflex suggesting a physiological explanation of behaviour patterns received a particularly warm welcome in philosophy and psychology as it provided a scientifically-based tool for a conceptual u-turn towards objectivism. This article looks into the ways these ideas contributed to the formation of the Soviet language theory, namely, to the sociological interpretation of language development and speech production presented in the pioneering works of Sergej M. Dobrogaev (1873–1952).

The frame/content theory of evolution of speech production

The species-specific organizational property of speech is a continual mouth open-close alternation, the two phases of which are subject to continual articulatory modulation. The cycle constitutes the syllable, and the open and closed phases are segments-vowels and consonants, respectively. The fact that segmental serial ordering errors in normal adults obey syllable structure constraints suggests that syllabic "frames" and segmental "content" elements are separately controlled in the speech production process. The frames may derive from cycles of mandibular oscillation present in humans from babbling onset, which are responsible for the open-close alternation. These communication-related frames perhaps first evolved when the ingestion-related cyclicities of mandibular oscillation (associated with mastication [chewing] sucking and licking) took on communicative significance as lipsmacks, tonguesmacks, and teeth chatters--displays that are prominent in many nonhuman primates. The new role of Broca's area and its surround in human vocal communication may have derived from its evolutionary history as the main cortical center for the control of ingestive processes. The frame and content components of speech may have subsequently evolved separate realizations within two general purpose primate motor control systems: (1) a motivation-related medial "intrinsic" system, including anterior cingulate cortex and the supplementary motor area, for self-generated behavior, formerly responsible for ancestral vocalization control and now also responsible for frames, and (2) a lateral "extrinsic" system, including Broca's area and surround, and Wernicke's area, specialized for response to external input (and therefore the emergent vocal learning capacity) and more responsible for content.     

The specificity of perceptual learning in speech processing

We conducted four experiments to investigate the specificity of perceptual adjustments made to unusual speech sounds. Dutch listeners heard a female talker produce an ambiguous fricative [?] (between [f] and [s]) in [f]- or [s]-biased lexical contexts. Listeners with [f]-biased exposure (e.g., [witlo?]; from witlof, "chicory"; witlos is meaningless) subsequently categorized more sounds on an [epsilonf]-[epsilons] continuum as [f] than did listeners with [s]-biased exposure. This occurred when the continuum was based on the exposure talker's speech (Experiment 1), and when the same test fricatives appeared after vowels spoken by novel female and male talkers (Experiments 1 and 2). When the continuum was made entirely from a novel talker's speech, there was no exposure effect (Experiment 3) unless fricatives from that talker had been spliced into the exposure talker's speech during exposure (Experiment 4). We conclude that perceptual learning about idiosyncratic speech is applied at a segmental level and is, under these exposure conditions, talker specific.     

Towards a computational theory of experience

A standing challenge for the science of mind is to account for the datum that every mind faces in the most immediate – that is, unmediated – fashion: its phenomenal experience. The complementary tasks of explaining what it means for a system to give rise to experience and what constitutes the content of experience (qualia) in computational terms are particularly challenging, given the multiple realizability of computation. In this paper, we identify a set of conditions that a computational theory must satisfy for it to constitute not just a sufficient but a necessary, and therefore naturalistic and intrinsic, explanation of qualia. We show that a common assumption behind many neurocomputational theories of the mind, according to which mind states can be formalized solely in terms of instantaneous vectors of activities of representational units such as neurons, does not meet the requisite conditions, in part because it relies on inactive units to shape presently experienced qualia and implies a homogeneous representation space, which is devoid of intrinsic structure. We then sketch a naturalistic computational theory of qualia, which posits that experience is realized by dynamical activity-space trajectories (rather than points) and that its richness is measured by the representational capacity of the trajectory space in which it unfolds.