Competition between whole-word and decomposed representations of English prefixed words


English aspiration is influenced by word structure: in general, a voiceless stop following s is unaspirated (des[t]royed), but it can be aspirated if a prefix-stem boundary intervenes (dis[th]rusts) (Baker et al. 2007). In a production study of 110 words prefixed with dis- or mis-, we show that even in prefixed words, there is variation (dis[k]laimersdis[kh]laimers), and that aspiration in such words is correlated with word and stem frequency. The more frequent the word, the less likely aspiration, but the more frequent the stem, the more likely aspiration. This contrasting frequency effect is characteristic of the type of competition Hay posits between whole-word lexical access and morphologically decomposed lexical access (Hay 2003): frequent words will tend to be accessed as wholes (and therefore behave as though there is no prefix-stem boundary), but frequent stems will encourage decomposed, prefix + stem access. In order to test whether there is active online competition, as opposed to simply frequency effects that are somehow lexicalized, we also conduct a priming experiment. We find that exposing participants to other prefixed words encourages them to aspirate target words, as compared to when they have been exposed to similar but non-prefixed words. These results provide evidence for active online competition.

This is a preview of subscription content, access via your institution.

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6
Fig. 7
Fig. 8
Fig. 9
Fig. 10
Fig. 11
Fig. 12


  1. 1.

    The Baker/Smith/Hawkins experiment, with only sixteen words (from fourteen lemmas) was not designed to investigate frequency effects systematically, but the researchers note that the pseudoprefixed words were more frequent and less semantically decomposable, and conjecture that they are treated as single units.

  2. 2.

    One additional participant was excluded for learning English at age 9, and another for equipment failure.

  3. 3.

    For example, Zue 1976 found an average VOT of about 25 msec. for English /t/ in utterances like /həˈstɑt/, a non-aspirating environment, and about 35 msec. in utterances like /həˈstɹɑt/, though this was still far less than the VOT for aspirated /t/ in an utterance like /həˈtɑk/, about 70 msec. Similarly, Klatt 1975 found that /t/’s VOT in English words beginning with /stV/ was on average 23 msec, and in words beginning /stɹV/ it was 35 msec (vs. 65 msec for words beginning with /tV/).

  4. 4.

    There were no random slopes included. Exploratory plots showed that slopes were very similar across participants for all variables except the interaction of word and stem frequency, but adding a random slope for that interaction produced non-convergence.

  5. 5.

    We added 1 to all frequencies before taking the log, to avoid taking the log of 0.

  6. 6.

    We used the vif() function in R’s car package (Fox and Weisberg 2010) to check the model’s degree of multicollinearity. VIF (variance inflation factor) values of 10 or larger are generally considered problematic. The largest VIF value in our model was 2.2, indicating that multicollinearity was not a serious problem for this model.

  7. 7.

    Using the plot_model() function in R’s sjPlot package (Lüdecke 2018), which relies on the ggplot2 package (Wickham 2016). We used the option type = “pred”, which gives predicted values, with discrete predictors “held constant at their proportions” (sjPlot documentation).

  8. 8.

    The highest VIF value found was 1.73, indicating that multicollinearity was not a serious problem in the model. (See footnote 7.)

  9. 9.

    In a pilot experiment, we used unproductively prefixed primes (conjugate) instead of totally unprefixed primes. Moreover, we did not separate the prime types into two blocks, but instead mixed unproductively and productively prefixed primes within both blocks, to test whether the immediately preceding prime influenced pronunciation. (We also did not include the meaning-judgement task described below.) The results showed that it was indeed possible for the same speaker, in the same experimental session, to pronounce the same word both ways (aspirated and unaspirated). Yet, there was no clear correlation between pronunciation and immediately preceding prime type. In this experiment we therefore strengthened the design: the two prime types are maximally different (productively prefixed vs. completely unprefixed) and occur in separate blocks, and a meaning-judgement task was added to ensure participants paid attention to the primes.

  10. 10.

    An additional three participants were recorded but excluded because their variety of English was non-U.S. (UK or South Asia).

  11. 11.

    The remaining participant reported beginning to learn English at age 15, which we take to be an error (perhaps 15 months was intended).

  12. 12.

    Median SUBTLEX-us (Brysbaert and New 2009) frequency count of target items was 41 (minimum 2, maximum 138). Median for prefixed primes was 26 (range: 0-362), and for prefixed fillers 5 (0-85). Median for unprefixed primes was 24.5 (range: 3-475), and for unprefixed fillers 14.5 (0-674). T-tests on log frequencies show the following significant frequency differences among sets: targets > {prefixed primes, unprefixed primes}, prefixed fillers < all others.

  13. 13.

    Discards was hard to categorize, because to discard has both a transparent sense (when playing cards, to remove a card from one’s hand) and an opaque sense (to throw away).

  14. 14.

    Experiment 1 produced similar results: as shown in Appendix A, mistaken and mistakes were aspirated by 12% and 0% of participants, whereas mistook was aspirated by 88%.

  15. 15.

    Alternatively, we could take as the dependent variable whether the aspirated token is in the prefix-prime block versus the unprefixed-prime block, such that the coefficient of interest is the intercept, and add the counterbalancing factor of block order as a fixed effect. The result is the same: no effect of block order, but the intercept is significantly different from zero (p = 0.04), indicating that the aspirated member of a perfect pair is significantly more likely to be in the prefixed block, across both block orders.

  16. 16.

    With only 20 target words, Experiment 2 was not designed to test frequency effects, and probably lacks the statistical power to do so. At a reviewer’s suggestion we added frequency factors to our model (log word CD count, log stem frequency, whether the stem exists as a freestanding word), singly and in a group, and did not find them to have a significant influence. We also tried adding word class to our model (only noun and verb, since there were so few adjectives and adverbs in this model), but it did not contribute significantly either. This is not to say that frequency and word class don’t matter, but our experimental design makes it unlikely that we could detect their effect.

  17. 17.

    p = 0.80 for block order, using a linear mixed-effects regression model with VOT change from first to second block as the dependent variable

  18. 18.

    Hanique and Ernestus (2012) argue that studies do not support a role for morphological decomposability in the fine details of phonetic reduction. (More recently, Plag and Ben Hedia (2018) do find effects of decomposability on the durations of un- and dis-, though not in- and -ly, in English corpus data.) We focus therefore on studies that, like our own, deal with variation between discrete phonological categories.

  19. 19.

    While there would be many choices of productively prefixed primes, such as misremember, there are few choices of unprefixed primes beginning with the same strings for comparison (about twenty-five in CELEX, depending on what one includes), and nearly all of those still suggest the negative meaning of dis- or mis- even if they are not productively prefixed, such as disappoint or dismantle.


  1. Akaike, H. (1998). Information theory and an extension of the maximum likelihood principle. In E. Parzen, K. Tanabe, & G. Kitagawa (Eds.), Selected papers of Hirotugu Akaike (Springer series in statistics) (pp. 199–213). New York: Springer. (3 May, 2019).

    Google Scholar 

  2. Audacity Team (1999). Audacity.

  3. Aylett, M., & Turk, A. (2004). The smooth signal redundancy hypothesis: a functional explanation for relationships between redundancy, prosodic prominence, and duration in spontaneous speech. Language and Speech, 47(1), 31–56.

    Article  Google Scholar 

  4. Baayen, R. H., Piepenbrock, R., & van Rijn, H. (1993). The CELEX lexical data base on CD-ROM. Linguistic Data Consortium.

    Google Scholar 

  5. Baker, R., Smith, R., & Hawkins, S. (2007). Phonetic differences between mis- and dis- in English prefixed and pseudo-prefixed words. In Proceedings of ICPhS XVI (pp. 553–556), Saarbrueken.

    Google Scholar 

  6. Baroni, M. (2001). The representation of prefixed forms in the Italian lexicon: evidence from the distribution of intervocalic [s] and [z] in northern Italian. In G. Booij & J. van Marle (Eds.), Yearbook of morphology 1999 (pp. 121–152). Dordrecht: Springer.

    Google Scholar 

  7. Bates, D., Maechler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects models using lme4. Journal of Statistical Software, 67(1), 1–48.

    Article  Google Scholar 

  8. Ben Hedia, S., & Plag, I. (2017). Gemination and degemination in English prefixation: phonetic evidence for morphological organization. Journal of Phonetics, 62, 34–49.

    Article  Google Scholar 

  9. Benua, L. (1997). Transderivational identity: phonological relations between words. Amherst: University of Massachusetts Press.

    Google Scholar 

  10. Boersma, P., & Weenink, D. (2006). Praat: doing phonetics by computer, version 4.4.

  11. Brysbaert, M., & New, B. (2009). Moving beyond Kucera and Francis: a critical evaluation of current word frequency norms and the introduction of a new and improved word frequency measure for American English. Behavior Research Methods, 41, 977–990.

    Article  Google Scholar 

  12. Bybee, J. (1985). Morphology: a study of the relation between meaning and form. Amsterdam: Benjamins.

    Google Scholar 

  13. Bybee, J. (1988). Morphology as lexical organization. In M. Hammond & M. Noonan (Eds.), Theoretical morphology: approaches in modern linguistics (pp. 119–142). San Diego: Academic Press.

    Google Scholar 

  14. Bybee, J. (2006). Frequency of use and the organization of language. London: Oxford University Press.

    Google Scholar 

  15. Bybee, J., & Scheibman, J. (1999). The effect of usage on degrees of constituency: the reduction of don’t in English. Linguistics, 37(4), 575–596.

    Article  Google Scholar 

  16. Caramazza, A., Laudanna, A., & Romani, C. (1988). Lexical access and inflectional morphology. Cognition, 28(3), 297–332.

    Article  Google Scholar 

  17. Chodroff, E., & Wilson, C. (2017). Structure in talker-specific phonetic realization: covariation of stop consonant VOT in American English. Journal of Phonetics, 61, 30–47.

    Article  Google Scholar 

  18. Fox, J., & Weisberg, H. S. (2010). An R companion to applied regression (2nd ed.). Thousand Oaks: Sage

    Google Scholar 

  19. Francis, A. L., Ciocca, V., & Ching Yu, J. M. (2003). Accuracy and variability of acoustic measures of voicing onset. The Journal of the Acoustical Society of America, 113(2), 1025–1032.

    Article  Google Scholar 

  20. Hanique, I., & Ernestus, M. (2012). The role of morphology in acoustic reduction. Lingue E Linguaggio, 11(2), 147–164.

    Article  Google Scholar 

  21. Hay, J. (2003). Causes and consequences of word structure. London: Routledge.

    Google Scholar 

  22. Johnson, K. (1997). Speech perception without speaker normalization: an exemplar model. In K. Johnson & J. W. Mullenix (Eds.), Talker variability in speech processing (pp. 145–165). San Diego: Academic Press.

    Google Scholar 

  23. Jurafsky, D., Bell, A., Gregory, M., & Raymond, W. D. (2001). Evidence from reduction in lexical production. In J. L. Bybee & P. Hopper (Eds.), Frequency and the emergence of linguistic structure (pp. 229–254). Amsterdam: Benjamins.

    Google Scholar 

  24. Kaye, A. S. (2005). Gemination in English. English Today, 21(2), 43–55.

    Article  Google Scholar 

  25. Klatt, D. H. (1975). Voice onset time, frication, and aspiration in word-initial consonant clusters. Journal of Speech, Language, and Hearing Research, 18(4), 686–706.

    Article  Google Scholar 

  26. Lüdecke, D. (2018). sjPlot - data visualization for statistics in social science.

  27. Ogden, R., Hawkins, S., House, J., Huckvale, M., Local, J., Carter, P., Dankovičová, J., & Heid, S. (2000). ProSynth: an integrated prosodic approach to device-independent, natural-sounding speech synthesis. Computer Speech & Language, 14(3), 177–210.

    Article  Google Scholar 

  28. Pagliuca, W. (1976). PRE-fixing. Manuscript. SUNY/Buffalo, ms.

  29. Pierrehumbert, J. (2001). Exemplar dynamics: word frequency, lenition, and contrast. In J. Bybee & P. Hopper (Eds.), Frequency effects and the emergence of linguistic structure (pp. 137–157). Amsterdam: Benjamins.

    Google Scholar 

  30. Plag, I., & Ben Hedia, S. (2018). The phonetics of newly derived words: testing the effect of morphological segmentability on affix duration. In S. Arndt-Lappe, A. Braun, C. Moulin, & E. Winter-Froemel (Eds.), Expanding the lexicon: linguistic innovation, morphological productivity (pp. 93–116). Berlin: de Gruyter.

    Google Scholar 

  31. R Core Team (2017). R: a language and environment for statistical computing. Vienna: R Foundation for Statistical Computing.

    Google Scholar 

  32. Raffelsiefen, R. (1999). Diagnostics for prosodic words revisited: the case of historically prefixed words in English. In T. A. Hall & U. Kleinhenz (Eds.), Studies on the phonological word (p. 133). Amsterdam: Benjamins.

    Google Scholar 

  33. Rosenfelder, I., Fruehwald, J., Evanini, K., & Yuan, J. (2011). FAVE (Forced Alignment and Vowel Extraction).

  34. Scicon R&D (2013). PCQuirerX.

  35. Seyfarth, S. (2014). Word informativity influences acoustic duration: effects of contextual predictability on lexical representation. Cognition, 133(1), 140–155.

    Article  Google Scholar 

  36. Shinners, P. (2000). PyGame.

  37. Smith, R., Baker, R., & Hawkins, S. (2012). Phonetic detail that distinguishes prefixed from pseudo-prefixed words. Journal of Phonetics, 40(5), 689–705.

    Article  Google Scholar 

  38. Sundara, M. (2005). Acoustic-phonetics of coronal stops: a cross-language study of Canadian English and Canadian French. The Journal of the Acoustical Society of America, 118(2), 1026–1037.

    Article  Google Scholar 

  39. Whalen, D. H., Best, C. T., & Irwin, J. R. (1997). Lexical effects in the perception and production of American English /p/ allophones. Journal of Phonetics, 25(4), 501–528.

    Article  Google Scholar 

  40. Wickham, H. (2016). ggplot2: elegant graphics for data analysis. New York: Springer.

    Google Scholar 

  41. Wurm, L. H. (1997). Auditory processing of prefixed English words is both continuous and decompositional. Journal of Memory and Language, 37(3), 438–461.

    Article  Google Scholar 

  42. Zue, V. W. (1976). Acoustic characteristics of stop consonants: a controlled study. Technical report 523. Lincoln Laboratory, Massachusetts Institute of Technology, ms.

  43. Zuraw, K. (2009). Frequency influences on rule application within and across words. Proceedings of CLS (Chicago Linguistic Society), 43, 283–309.

    Google Scholar 

  44. Zuraw, K., & Peperkamp, S. (2015). Aspiration and the gradient structure of English prefixed words. In The Scottish Consortium for ICPhS 2015 (Ed.), Proceedings of the 18th International Congress of Phonetic Sciences. Paper number 0382, pp. 1–5). Glasgow, UK: The University of Glasgow.

Download references


This work was supported by Faculty Research grants from the UCLA Academic Senate’s Committee on Research, as well as by the Agence Nationale de la Recherche (ANR-17-CE28-0007-01, ANR-17-EURE-0017). We are grateful to members of the UCLA Phonology seminar and the audience of the 2018 DGfS workshop on ‘Variation and phonetic detail in spoken morphology’ for comments and discussion. We especially thank the many undergraduate research assistants who processed sound files, coded data, and ran participants: Scarlet Yejin Cho, Peter Hee Hwan, Scott Boegeman, Abigail Carlson, Evan Davis-Palley, Liam Donohue, Jennifer Gethers, Patrick Ryan Kelly, Jessie Ng, Amy Tang, Zhuoren Tong, Zachary Thomas, Stephanie Wang, Chris Yang, and Michael Zandona.

Author information



Corresponding author

Correspondence to Kie Zuraw.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


Appendix A: Materials for Experiment 1, with results

Target words:

  mean VOT (msec) standard deviation number unaspirated number aspirated number unsure
discards 39.4 22.1 10 4 2
disclaimers 61.3 26.2 5 10 1
disclosing 50.9 18.1 5 7 4
disclosures 47.9 17.3 6 5 4
disco 24.8 8.3 16 0 0
discolorations 60.2 19.3 2 12 2
discolored 57.7 22.8 5 10 1
discoloring 69.3 25.4 3 12 1
discomfort 51.5 19.2 3 13 0
disconcerting 46.1 15.8 1 12 1
disconcerts 55.4 16.1 0 11 2
disconnect 48.4 13.4 1 13 2
discontent 48.9 18.8 2 14 0
discontinued 45.2 11.8 1 14 1
discontinuity 60.7 20.8 0 13 1

Filler words:

collapse considerately impairing innovate reduce replaceable
collapsed consignment impaling inquired reductions replenished
collated consistencies imparted inscription refinement replenishments
collating consolidate impatient insecticide refiner replica
collectors conspired impeached insomnia reflector replicas
collide constrain imperatives inspects refresher reported
collided constraining imported inspiring refrigerators represses
combined constricted imposed installed refunded republicanism
commanders constrictions impound instilling refute repugnance
commandment constructed impoverished insulated refuting repulsed
commissary constructive impregnate insulted regained requesting
commissioned construing imprisoned insure regalia required
commodities consulates improper intention regarded rerunning
communion consultant imprudence intern regatta researching
commuted consultants inaccuracies interned regattas resemblance
commuting consummately incisors interning regress resembled
compact contacted inciting internment regressed reserving
companion contagiously include intimating regressing resettlement
compartment container incompetency intruded regurgitate resided
compelled contaminated incongruous invade regurgitates residual
competitor contemptible inconsistencies invading rehearsals resilience
compile contended incurs invariably reinsuring resistances
compiles contestants indented inversion rejoice resistors
compiling contextual indestructibility invested rejoicings resort

Appendix B: Materials for Experiment 2

Target words and associated primes:

target word prefixed prime unprefixed prime
discards uproots laments
disclaimers coequals placebos
disclosing enriching policing
discolored unfailing harmonic
discomfort injustice momentum

Prefixed fillers:

befriend enraged indelicately rearmament unfathomably
begrudging enrapturing inelegantly redeploy unflinching
coexist enriched ineligibly refashioning unleashes
coexistence enriches ingratitude reforest unlikelihood
cohabit enshrouding inhabitant refreshes unravels
derails ensnare insubordination rehabilitation unreasonably
dethrone enthroning insufferable reinsure unseemly
embittered envenoming maladjustment reran unsightly
embodying foreshadow prearrange subdivide unswerving
emboldened foreshadowing predigest transforms untiringly
enable foreshortening prerecord unalterably unutterable
encircle forewarn presuppose unbeliever unvarnished
endanger immoderately reacts undoubted unzip
endangers immodesty readjustment unearthly unzipped
enfolding inadequately reaffirm unending upholds
enlarges indecency reappear unfalteringly withhold

Unprefixed fillers:

ballooning cements horizons monastic pollutes
battalions cocoons hyphenation negates questionnaire
benevolently cultivation idyllic neglect rapacious
bombastic facetious imagines nomadic rumination
bravado fallacious lamenting nominee salacious

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Zuraw, K., Lin, I., Yang, M. et al. Competition between whole-word and decomposed representations of English prefixed words. Morphology 31, 201–237 (2021).

Download citation


  • English
  • Aspiration
  • Priming