@article {516, title = {What are the mental representations of speech segments during speech preparation, as revealed by self-monitoring for speech errors?}, year = {2022}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Utrecht, The Netherlands}, abstract = {Segmental speech errors can be detected by self-monitoring both before and after speech is initiated. This allows us to answer four questions about speech preparation: (1) What is the delay between self-monitoring internal and external speech? (2) What are the representations of speech sounds on which self-monitoring focuses during {\textquotedblleft}early{\textquotedblright} and {\textquotedblleft}late{\textquotedblright} error detection? (3) Why are some speech errors detected by self-monitoring before and others after speech initiation, and others again not at all?(4) What is the role of cognitive control in self-monitoring for speech errors? For answering those questions we use responses from 6 of our earlier SLIP experiments. We find that (1) the delay between {\textquotedblleft}early{\textquotedblright} and {\textquotedblleft}late{\textquotedblright} error detection is roughly 4 segments or nearly 500 ms. (2) Speech sounds are represented very differently before and after speech initiation. This is not predicted from articulatory phonology but supports the proposal that in internal speech sounds are represented as targets in auditory perceptual space. (3) Phonetic contrast between segments is a major factor determining whether speech errors are detected internally, externally or not at all. (4) Degree of conflict between competing items during speech preparation controls frequency of speech errors, but not necessarily of error detection. }, author = {Sieb Nooteboom and Hugo Quen{\'e}} } @article {495, title = {Phonetic contrast between error and correct target is a major factor in determining whether segmental errors are detected "early", "late", or not at all}, year = {2021}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {online}, abstract = {In self-monitoring for speech errors, some segmental speech errors are detected "early", i.e. before articulation is initiated, and others "late", i.e. after articulation has started, and others again are detected not at all. What causes these differences in detecting speech errors? We hypothesize that detection of errors that are similar to the correct target takes longer than detection of errors that are dissimilar. We also hypothesize that the time available for error detection in internal speech is limited. If this time is exceeded before an error is detected, self-monitoring gets a second chance during or after articulation. If also the time available for this second stage of error detection is exceeded, then the error remains undetected. From these hypotheses we predict that relatively more "dissimilar" than "similar" segmental errors are detected before articulation, and also that relatively more "similar" than "dissimilar" errors remain undetected. To test these hypotheses, we have assessed, in data obtained in two earlier published SLIP experiments, the relative frequencies of repaired and unrepaired segmental speech errors that differ systematically in strength of phonetic contrast with the correct target. Results show that indeed a major factor affecting the distribution of errors over "early", "late" and "undetected" is the strength of phonetic contrast between two competing response candidates.}, author = {Sieb Nooteboom and Hugo Quen{\'e}} } @article {457, title = {Testing a conflict-based theory of self-monitoring for speech errors}, year = {2019}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Amsterdam, The Netherlands}, abstract = {This paper reports an experimental test of a theory of self-monitoring proposed by Nozari, Dell and Schwartz (2011). The theory presupposes that multiple items generated by the speech production system may be active simultaneously, correct and incorrect items competing for the same slot. In case of error, conflict information is passed on to an executive control center, leading to detection. When the overall conflict in the system increases, distinguishing between correct and error trial becomes more difficult, which in turn may result in (1) more errors being made, and (2) fewer errors being detected. We test this theory with data obtained in two experiments eliciting segmental speech errors in Dutch CVC CVC word pairs, as reported in Nooteboom \& Quen{\'e} (accepted for publication). The two experiments accidentally differed in overall conflict. All of the stimuli in the first experiment, with least overall conflict, were also used in the second experiment. There overall conflict was much higher because of the inclusion of a condition eliciting errors against the relatively weak voiced-voiceless feature in initial stop consonants, leading to more errors against this feature, and to fewer errors being detected. The crucial comparison is between the conditions with identical stimuli. There we find significantly fewer errors being detected in the second than in the first experiment. This supports a conflict-based theory of speech error detection. References Nooteboom S.G. \& Quen{\'e}. H. (accepted for publication). Repairing segmental speech errors. Competition as a source of repairs. Journal of Memory and Language. Nozari, N., Dell, G., \& Schwartz, M. (2011). Is comprehension necessary for error detection? A conflict-based account of monitoring in speech production. Cognitive Psychology 63, 1{\textendash}33. }, author = {Sieb Nooteboom and Hugo Quen{\'e}} } @article {445, title = {Repairing segmental speech errors: The role of competing lexical items}, year = {2018}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Amsterdam, The Netherlands}, abstract = {We propose that in repairing segmental speech errors {\textquotedblleft}repairs{\textquotedblright} stem from word forms that during speech preparation and self-monitoring compete with the word form selected for being spoken. Activation of these potential repairs decreases during the time lag (500 ms) between detection in internal and in overt speech. Recently, it was demonstrated that repaired speech errors can be classified as detected in internal or in overt speech. A re-analysis of data obtained in two experiments eliciting speech errors and their repairs, shows that: (1) Error-to-interruption times are shorter after single elicited errors than after non-elicited and multiple (together "other") errors. (2) Single elicited errors are relatively more often detected in internal speech than "other" errors. (3) The correct word form is the most frequent form used as repair, but less frequently after detection in overt speech than after detection in internal speech. (4) Interruption-to-repair times are shorter for single elicited than for other errors, but less so after detection in overt speech. These findings support the new theory of repairing. }, author = {Sieb Nooteboom and Hugo Quen{\'e}} } @article {21, title = {Plan, Scan, Speak, Detect, Stop, Repair}, year = {2017}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Amsterdam, The Netherlands}, abstract = {

In this paper we develop some predictions on interactional segmental speech errors and their repairs, based on a computational model of self-monitoring proposed by Hartsuiker and Kolk (2001) and modified by Nooteboom and Quené (2017). These predictions were tested in a four-word tongue twister experiment eliciting such errors and their repairs, in word initial and word medial position.

Findings are: (1) The statistical distributions of error-to-cutoff times, although truncated close to 0 ms, are nearly complete for both initial and medial consonant errors. This implies that against prediction interruption of the speaking process after internal error detection takes more time than speech initiation. (2) The distributions of cutoff-to-repair times are censored at 0 ms, but cutoff-to-repair times are longer for medial than initial consonants, showing that against prediction repairing takes more time for medial than for initial errors. (3) Detection rate is much lower for medial than for initial consonants and decreases with position of the misspoken word in the tongue twister sequence. This probably reflects predicted variation in selective attention.

}, author = {Sieb Nooteboom and Hugo Quen{\'e}} } @article {16, title = {Self-monitoring for speech errors: Two-stage detection and repair with and without auditory feedback}, year = {2016}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Utrecht, Nederland}, abstract = {Two experiments are reported, eliciting segmental speech errors and self- repairs. Error frequencies, detection frequencies, error-to-cutoff times and cutoff-to-repair times were assessed with and without auditory feedback, for errors against four types of segmental oppositions. Main hypotheses are (a) prearticulatory and postarticulatory detection of errors is reflected in a bimodal distribution of error-to-cutoff times; (b) after postarticulatory error detection repairs need to be planned in a time-consuming way, but not after prearticulatory detection; (c) postarticulatory error detection depends on auditory feedback. Results confirm hypotheses (a) and (b) but not (c). Internal and external detection are temporally separated by some 500 ms on average, fast and slow repairs by some 700 ms. Error detection does not depend on audition. This seems self-evident for prearticulatory but not for postarticulatory error detection. Theoretical implications of these findings are discussed. }, author = {Sieb Nooteboom and Hugo Quen{\'e}} } @article {47, title = {Word onset effect in interactional speech errors: spontaneous speech versus elicitation}, year = {2014}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Utrecht, The Netherlands}, abstract = {

Interactional speech errors are speech errors that have an obvious source in the immediate environment. It has been observed that segmental interactional speech errors are more frequent in word onsets than in other positions. Earlier we have shown that in spontaneous speech this so-called "word onset effect" in interactional errors in spontaneous speech may be fully explained by the number of opportunities segments in different positions in the word have for interaction with other segments in the immediate environment. However, there is experimental evidence for a real word onset effect in elicited interactional speech errors.

We will report a tongue twister experiment set up to investigate interactional substitutions of consonants under different elicitation conditions. Results show a real and considerable word-onset effect if the interacting consonants share both word-onset position and pre-stress position as compared to a condition in which they share a word medial position plus pre-stress position. However, when the interacting consonants share neither position in the word nor pre-stress position, error frequency is completely determined by numbers of opportunities. We conclude that the word onset effect is limited to experiments that successfully elicit interactions by selective activation of segments.

}, author = {Sieb Nooteboom and Hugo Quen{\'e}} } @article {61, title = {Waar komt de dominantie van woordbegin-medeklinkers in klankversprekingen vandaan?}, year = {2013}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Utrecht, The Netherlands}, abstract = {

Beginmedeklinkers van woordvormen zijn relatief veel vaker bij klankversprekingen betrokken dan medeklinkers in andere posities. Shattuck-Hufnagel (1987, 1992, 2011) verklaart dat door aan te nemen dat bij het klaarzetten van woordvormen om uitgesproken te worden beginklanken van woorden apart van de rest behandeld worden. Deze complicatie is onnodig wanneer we een andere verklaring vinden voor die overheersing van woordbeginmedeklinkers. Die zoeken wij in de fonotactische structuur van de taal. We hebben een analyse gemaakt van een corpus klankversprekingen in spontaan Nederlands, en van de fonotactische structuur van de woordreeksen waarin die versprekingen zijn gemaakt. Het blijkt dat de relatieve frequentie van klankversprekingen in verschillende posities binnen woorden voorspeld kan worden uit de fonotactische mogelijkheden die iedere klank heeft om versproken te worden. We hebben ook een experiment gedaan waarin klankversprekingen zijn uitgelokt in CVC-woorden, in de posities van C1, V en C2. De resultaten laten zien dat C1 en C2 even vaak versproken worden, maar de klinker minder vaak. Dat laatste kan verklaard worden door aan te nemen dat op elkaar lijkende klinkers meer verschillen dan op elkaar lijkende medeklinkers. Het is onnodig om te veronderstellen dat in de spraakvoorbereiding woordbeginklanken apart behandeld worden.\ 

}, author = {Sieb Nooteboom and Hugo Quen{\'e}} } @article {87, title = {Parallellen tussen zelfbewaking bij het spreken en identificatie van spraaksegmenten door luisteraars}, year = {2012}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Utrecht, Nederland}, abstract = {

In deze bijdrage rapporteren we over een experiment waarin we aspecten van de zelfbewaking bij het spreken op een indirecte manier onderzoeken. Luisteraars hebben 291 korte spraakfragmenten ge{\"\i}dentificeerd die beginmedeklinkers bevatten, gesneden uit segmentele versprekingen zoals uitgelokt in eerdere experimenten. Voor ieder versproken segment is er een correct controlesegment van dezelfde spreker. Segmentele versprekingen waren {\'o}f niet ontdekt door de spreker, {\'o}f vroeg ontdekt zoals blijkt uit een interruptie (wee... beter weten), {\'o}f laat ontdekt (weter beten... beter weten). Aantal misidentificaties en reactietijden zijn bepaald. Misidentificatie is iedere afwijking van een eerdere auditieve transcriptie. Aantal misidentificaties is laag (3.1 \%), maar toch significant hoger voor versproken dan voor controlesegmenten, en significant hoger voor laat dan voor vroeg gedetecteerde versprekingen. Reactietijden zijn gemeten voor spraakfragmenten die door de luisteraars unaniem correct waren ge{\"\i}dentificeerd. Deze reactietijden verraden desondanks effecten van vari{\"e}rende perceptieve ambigu{\"\i}teit. Vroeg gedetecteerde versprekingen en correcte controlefragmenten hebben samen de kortste reactietijden, laat gedetecteerde versprekingen hebben significant langere reactietijden. We speculeren dat vroeg gedetecteerde versprekingen zelfbewaking van innerlijke spraak reflecteren, met als criterium afwijking van de correcte doelklank, en laat gedetecteerde versprekingen zelfbewaking van de articulatie met als criterium conflicterende articulatiebewegingen.

}, author = {Sieb Nooteboom and Hugo Quen{\'e}} } @article {92, title = {Weter beten: Verwisselingen overheersen in segmentele versprekingen}, year = {2011}, publisher = {Nederlandse Vereniging voor Fonetische Wetenschappen}, address = {Utrecht, The Netherlands}, abstract = {

Het zogenaamde "scan-copier" model voor de seri{\"e}le ordening van spraaksegmenten van Shattuck-Hufnagel (1983) voorspelt, in strijd met wat we in collecties versprekingen aantreffen, dat er meer segmentele verwisselingen dan anticipaties en perseveraties zijn (voorspelling 1). Recent onderzoek van Goldstein et al. (2007) en van McMillan \& Corley (2010) laat zien dat segmentele versprekingen gradueel zijn en het resultaat kunnen zijn van parti{\"e}le en simultane activatie van segmenten die strijden om dezelfde positie. Dit leidt tot voorspelling 2: Responsietijden voor niet verbeterde versprekingen zijn langer dan die voor wel verbeterde versprekingen. Beide voorspellingen blijken bij toetsing op te gaan. Dit steunt zowel het scan-copier model van Shattuck-Hufnagel als het idee van Goldstein c.s. dat versprekingen gradueel zijn en segmenten kunnen concurreren voor dezelfde positie.

}, author = {Sieb Nooteboom and Hugo Quen{\'e}} }