Evidence for the Modulation of Sub-Lexical Processing in Go No-Go Naming: The Elimination of the Frequency ? Regularity Interaction

Abstract

The Frequency (high vs. low) × Regularity (regular vs. exception) interaction found on naming response times is often taken as evidence for parallel processing of sub-lexical and lexical systems. Using a Go/No-go naming task, we investigated the effect of nonword versus pseudohomophone foils on sub-lexical processing and the subsequent Frequency × Regularity interaction. We ran two experiments: (1) a Go/No-go naming task with nonword foils (e.g., bint) and (2) a Go/No-go naming task with pseudohomophone foils (e.g., pynt). Experiment 1 replicated the Frequency × Regularity interaction on naming response times supporting the notion of parallel sub-lexical and lexical processing. Experiment 2 eliminated the Frequency × Regularity interaction providing evidence for the modulation of sub-lexical information. These results indicate that using pseudohomophones in the Go/No-go naming task minimized information provided from sub-lexical processing and maximized information provided from the lexical system.[PUBLICATION ABSTRACT]

J Psycholinguist Res (2011) 40:367378 DOI 10.1007/s10936-011-9174-2

Evidence for the Modulation of Sub-Lexical Processing

in Go No-Go Naming: The Elimination of the Frequency Regularity Interaction

Jacqueline Cummine Josee Amyotte

Brent Pancheshen Brea Chouinard

Published online: 18 September 2011 Springer Science+Business Media, LLC 2011

Abstract The Frequency (high vs. low) Regularity (regular vs. exception) interaction

found on naming response times is often taken as evidence for parallel processing of sub-lexical and lexical systems. Using a Go/No-go naming task, we investigated the effect of non-word versus pseudohomophone foils on sub-lexical processing and the subsequent Frequency

Regularity interaction. We ran two experiments: (1) a Go/No-go naming task with nonword foils (e.g., bint) and (2) a Go/No-go naming task with pseudohomophone foils (e.g., pynt). Experiment 1 replicated the Frequency Regularity interaction on naming response times

supporting the notion of parallel sub-lexical and lexical processing. Experiment 2 eliminated

the Frequency Regularity interaction providing evidence for the modulation of sub-lexical

information. These results indicate that using pseudohomophones in the Go/No-go naming task minimized information provided from sub-lexical processing and maximized information provided from the lexical system.

Keywords Sub-lexical de-emphasis Strategic control Go/No-go naming Regularity

Frequency Pseudohomophones

J. Cummine (B) J. Amyotte B. Chouinard

Department of Speech Pathology and Audiology, Faculty of Rehabilitation Medicine, University of Alberta, 2-70 Corbett Hall, 8205 114 St., Edmonton, AB, T6G 2G4, Canada e-mail: [email protected]

B. PancheshenDepartment of Linguistics, University of Alberta, Edmonton, AB, Canada

123

368 J Psycholinguist Res (2011) 40:367378

Introduction

The extent to which readers can emphasize processing along the sub-lexical or lexical routes in the English language has been of considerable debate (Lupker et al. 1997; Monsell et al. 1992; Visser and Besner 2001). Strategic control of lexical and sub-lexical reading systems, whether conscious or unconscious, has been investigated by manipulating the presentation format of high and low frequency regular and exception words (e.g., mixing stimuli, repetition of stimuli; Besner et al. 2010; Kinoshita et al. 2004; Lupker et al. 1997; Monsell et al. 1992). For example, high frequency exception words are named significantly faster when presented in a pure block format compared to a mixed block with nonwords. Presumably, this is because a pure block presentation format is a condition in which participants can strategically utilize information from the lexical system by de-emphasizing (Monsell et al. 1992) or setting a time criterion on the sub-lexical processing system (Lupker et al. 1997). While these results have been inuential in furthering our understanding of lexical and sub-lexical processing, the notion of strategic route emphasis/control remains ambiguous due to the dearth of research in the area. This paper furthers the investigation of strategic control of lexical and sub-lexical reading systems by changing the foils in a Go/No-go naming task in order to encourage information processing from the lexical and/or sub-lexical system. Subsequent frequency and regularity effects can be used as evidence of lexical and sub-lexical information processing, respectively. Utilizing a Go/No-go naming task is a novel approach that provides a direct evaluation of strategic emphasis on lexical and sub-lexical processing.

A Go/No-go naming task requires that participants name aloud a stimulus only if it meets the requirement of being a word. While this task has proven useful in investigations of word frequency and regularity (Hino and Lupker 1998, 2000), the utility of the Go/No-go naming task in emphasizing reliance on lexical and/or sub-lexical information has been overlooked. By changing the foils to be either nonwords or pseudohomophones we can investigate the strategic reliance on lexical and/or sub-lexical systems. Specifically, a Go/No-go task with nonwords can be successfully completed when participants use information from both sub-lexical and lexical systems to make a go/no go decision because nonwords do not have whole-word (lexical) representations and do not sound like real words when they are processed via the sub-lexical system. The use of both systems should be evidenced by a significant Frequency Regularity interaction, which is the nding that low frequency exception words

(e.g., sieve) are named significantly slower than low frequency regular words (e.g., swell), whereas higher frequency exception and regular words are named relatively faster and do not demonstrate the same magnitude of effect (Besner and Smith 1992; Lupker et al. 1997; Paap and Noel 1991). The Frequency Regularity interaction is often taken as evidence for the

contribution of both sub-lexical and lexical systems during reading aloud (Besner and Smith 1992; Lupker et al. 1997; Paap and Noel 1991). In contrast, a Go/No-go task with pseudohomophones encourages participants to rely on or emphasize information from the lexical system to make a decision given that sub-lexical processing of pseudohomophones would encourage a go response because they sound like real words when sounded out. If participants de-emphasize the sub-lexical route we should see the elimination of the Frequency

Regularity interaction and a response advantage for low frequency exception words in the Go/No-go naming task with pseudohomophones when compared to the Go/No-go naming task with nonwords. To test these hypotheses, we ran two experiments: a Go/No-go naming task with nonwords and a Go/No-go naming task with pseudohomophones. The high and low frequency regular and exception words were identical in both experiments and only the foils were changed.

123

J Psycholinguist Res (2011) 40:367378 369

Materials and Methods

Experiment 1. Go/No-Go Naming with Nonwords

Participants

A total of 22 undergraduate students from the University of Alberta participated in this experiment for course credit. Inclusion criteria consisted of normal or corrected normal vision and English as a rst language. The subjects consent was obtained according to the Declaration of Helsinki (1996) and the experiment was performed in compliance with the relevant laws and institutional guidelines and was approved by the University of Alberta Health Research Ethics Board.

Stimuli

The stimuli consisted of regular words (e.g., swell), exception words (e.g., sieve), and non-words (e.g., bint; see Appendix A). There were 27 regular words (13 high frequency, mean Kucera and Francis (KF) word frequency=683, 14 low frequency, mean KF word

frequency=20), 27 exception words (13 high frequency, mean KF word frequency=716;

14 low frequency, mean KF word frequency=40), and 23 nonwords. Stimuli were matched

for onset phoneme, length, bigram sum, frequency, phonological neighbourhood, and ortho-

graphic neighbourhood.1

Materials

Stimuli were presented on a computer monitor using EPrime software (Psychology Software Tools, Inc., http://www.pstnet.com

Web End =http://www.pstnet.com ). Voice onset was coded via a microphone and the experimenter used a button response to code accuracy on each trial.

Procedure

After giving written consent, participants were tested individually in a normally lit room. Letter strings were presented randomly to the centre of a computer screen. Participants were instructed to name the stimulus aloud, if the letter string spelled a real word, as quickly and accurately as possible (totaling 70% of the presented stimuli). Participants were given 1,800ms to name the stimuli. The experimenter used a mouse button click to code correct (left button) and incorrect (right button) responses on all word trials.

1 Using a database dictionary resource (MRC Psycholinguistic Database: Machine Usable Dictionary. Version2.00, http://www.psy.uwa.edu.au/mrcdatabase/mrc2.html

Web End =http://www.psy.uwa.edu.au/mrcdatabase/mrc2.html ), which provides a count of the number of occurrences of particular linguistic properties we evaluated our stimuli on several characteristics. When comparing high (and low) frequency regular and exception words, the stimuli did not differ on Bi-gram sum (sum of frequencies for consecutive bi-grams; High frequency, p = .963; Low frequency, p = .205; as dened by Balota

et al. 2007), written frequency (counts per million; High frequency p = .829; Low frequency p = .057),

length (number of letters in a word; High frequency p = .789; Low frequency p = .304), phonological

neighbourhood (number of correctly sounding words that can be made by replacing one letter at a time; High frequency p = .770; Low frequency p = .194) or orthographic neighbourhood (number of correctly spelled

words that can be made by replacing one letter at a time; High frequency p = .392; Low frequency p = .054)

characteristics.

123

370 J Psycholinguist Res (2011) 40:367378

Table 1 Mean response times (SE) for stimuli in experiment 1

Go No-Go with Nonwords

Exception words Regular words

High frequency 667.38 (19.72) 672.04 (21.74)

Low frequency 790.13 (24.58) 716.47 (21.37)

Results

Only correct responses were included in the subsequent analyses (Table 1). In keeping with previous literature, trials were removed if reaction times were <250, >1,800ms or where a voice key error was made (Hino and Lupker 1998, 2000). The mean accuracy rates were as follows: regular words=94%, exception words=90%, nonwords=86%. A 2 (Frequency)

2 (Regularity) repeated measures ANOVA was conducted for by-subject and by-item

analyses.

By-Subject Analyses

There was a significant Frequency effect, F(1, 21) = 29.134, p < .001, where high

frequency words (Mean = 669.711 ms 19.865 ms) were responded to faster than low

frequency words (Mean = 753.304 ms 21.452 ms; generalized 2 = .148, partial

2 = .581; see Olijnik and Algina (2003) for calculation of generalized 2 for repeated

measures). There was a significant Regularity effect, F(1, 21) = 9.332, p = .006, where

regular words (Mean = 694.258 ms 19.904 ms) were responded to faster than excep

tion words (Mean = 728.757 ms 20.062 ms; generalized 2 = .029, partial 2 =

.308). There was a significant overadditive interaction between Frequency and Regularity, F(1, 21) = 17.909, p < .001 (generalized 2 = .037, partial 2 = .460; see Fig. 1).

The effect of Regularity was significantly greater for low frequency words with a difference of 73.66ms between regular words and exception words in comparison to high frequency words, which only showed a difference of 4.662ms. Response times to high frequency regular words (Mean=672.042ms 21.740 ms) were not significantly faster than high frequency

exception words (Mean = 667.380 ms 19.729 ms), t(21) .387, p > .05, and low fre

quency regular words (Mean = 716.474 ms 21.367 ms) were significantly faster than low

frequency exception words (Mean = 790.134 ms 24.585 ms), t(21) = 4.393, p < .001

(uncorrected).

By-Item Analyses

The item analyses show the same pattern of results as the subject analyses. There was a significant Frequency effect, F(1, 12) = 13.079, p = .004, a Regularity effect that

approached significance, F(1, 12) = 3.537, p = .085, and a significant overadditive inter

action between Frequency and Regularity, F(1, 12) = 12.76, p = .004. The effect of

Regularity was significantly greater for low frequency words with a difference of 64.23ms between regular words and exception words in comparison to high frequency words, which only showed a difference of 6.39ms. Response times to high frequency regular words (Mean = 666.95 ms10.78 ms) were not significantly faster than high frequency exception

123

J Psycholinguist Res (2011) 40:367378 371

Reaction time (ms)

Fig. 1 Go/No-go naming with Nonwords: significant frequency (high vs. low) regularity (regular words

vs. exception words) interaction on behavioural reaction time

words (Mean = 660.56 ms 10.12 ms), t(13) = .657, p > .05, and low frequency reg

ular words (Mean = 723.72 ms 19.91 ms) were significantly faster than low frequency

exception words (Mean = 787.95 ms 32.10 ms), t(12) = 2.765, p = .017 (uncorrected). Experiment 2. Go/No-Go Naming with Pseudohomophones

Participants

A total of 37 undergraduate students from the University of Alberta participated in this experiment for course credit. None had participated in Experiment 1. Inclusion criteria and ethical approval were identical to Experiment 1.

Stimuli

The stimuli were identical to those used in Experiment 1 except that 23 pseudohomophones (e.g., pynt; 12 high frequency, mean KF word frequency=225, 11 low frequency, mean KF

word frequency=23) were used instead of nonwords.

Materials and Procedure

The materials and procedures were identical to Experiment 1.

Results

Only correct responses and trials in which reaction times were >250ms were included in the subsequent analyses (see Table 2). The mean accuracy rates were as follows: regular words=96%, exception words=93%, pseudohomophones=89%. A 2 (Frequency) 2

(Regularity) repeated measures ANOVA was conducted for by-subject and by-item analyses.

Exception

Regular

High

Low

Frequency

123

372 J Psycholinguist Res (2011) 40:367378

Table 2 Mean response times (SE) for stimuli in experiment 2

Go No-Go with Pseudohomophones

Exception words Regular words

High frequency 625.45 (16.47) 607.19 (16.68)

Low frequency 701.28 (23.18) 686.07 (23.75)

Exception

Regular

Fig. 2 Go/No-go naming with Pseudohomophones: significant frequency (high vs. low) effect on behavioural reaction time

By-Subject Analyses

There was a significant Frequency effect, F(1, 35)=35.038, p < .001, where high frequency

words (Mean = 616.321 ms 15.863 ms) were responded to faster than low frequency

words (Mean = 693.678 ms 22.075 ms; generalized 2 = .094, partial 2 = .500). There

was no significant Regularity effect, F(1, 35) = 2.334, p > .05, (regular words Mean =

646.631 ms18.616 ms and exception words Mean = 663.368 ms19.157 ms). There was

no significant interaction between Frequency and Regularity, F(1, 35) = .044, p > .05 (see

Fig. 2). The Regularity effect was not significantly greater for low frequency words, with a difference of 15.214ms between regular words and exception words, in comparison to high frequency words, which showed a difference of 18.262ms. Response times to high frequency regular words (Mean = 607.190 ms 16.677 ms) were not significantly faster than high fre

quency exception words (Mean = 625.452 ms 16.466 ms), t(35) = 1.904, p > .05, and

low frequency regular words (Mean = 686.071 ms 23.748 ms) were not significantly

faster than low frequency exception words (Mean = 701.285 ms 23.177 ms) t(35) =

.956, p > .05.

By-Item Analyses

Again, the item analyses show the same pattern of results as the subject analyses. There was a significant Frequency effect, F(1, 12)=17.107, p = .001, no Regularity effect,

Reaction time (ms)

High

Low

Frequency

123

J Psycholinguist Res (2011) 40:367378 373

F(1, 12)=2.708, p > .05, and importantly, there was no interaction between Frequency

and Regularity, F(1, 12)=.668, p > .05. Response times to high frequency regular words

(Mean=608.08ms 4.89ms) were not significantly faster than high frequency exception

words (Mean=625.61ms 12.72ms), t(13)=1.253, p > .05, and low frequency regu

lar words (Mean=681.12ms 12.85ms) were not significantly faster than low frequency

exception words (Mean=709.16ms 29.84ms), t(12)=1.272, p > .05.

To test the last hypothesis that the Go/No-go naming task with pseudohomophones would serve to decrease the reaction times for naming low frequency exception words, we ran an independent samples t test on the low frequency exception words from experiment 1 and2. As expected, the low frequency exception words in the Go/No-go naming with pseudohomophones (701.285ms 23.177ms) were named faster than the low frequency excep

tion words in the Go/No-go naming with nonwords (790.134ms 24.584ms). Using an

adjusted p-value based on four comparisons, this difference was marginally significant at p = .0125, t(56) = 2.513, p = .015 (Cohens d=.695).2 At the adjusted p-level, all other

pair-wise comparisons were nonsignificant (p > .0125).

Discussion

Our study provides the rst demonstration that (1) manipulating the foils used in a Go/No-go naming task inuences the strategies with which participants make responses to the presented stimuli and (2) that the Frequency Regularity interaction can be eliminated in a naming

task. Specifically, the results provide evidence that the Go/No-go naming task with pseudohomophones encouraged participants to emphasize information processing from the lexical system and de-emphasize information processing from the sub-lexical system. In contrast, the Go/No-go naming task with nonwords promoted the use of information processing from both sub-lexical and lexical systems. When combined, our results support the notion that there is separability between sub-lexical and lexical processing systems and that Frequency and Regularity effects can be used to make inferences about processing along these streams.

To our knowledge, this is the rst demonstration of a naming task in which Frequency and Regularity do not interact. While previous work has demonstrated that the Frequency

Regularity interaction can be reduced as a result of repetition effects (Visser and Besner

2001), we provide a clear demonstration that sub-lexical phonological information can also be reduced by presenting stimuli that encourage strategic reliance on lexical processing. The Frequency Regularity effect is often taken as evidence of the contribution of both lexical

and sub-lexical information (Besner and Smith 1992; Coltheart et al. 2001; Paap and Noel 1991). Specifically, it is argued that high frequency exception words can be processed fast enough via the lexical route so that naming occurs before the sub-lexical route can create a competing and incorrect phonological code. Low frequency exception words, on the other hand, take longer to process and thus the sub-lexical route can produce a competing phono-logical representation that must ultimately be resolved with the lexical pronunciation prior to speech output.

Two accounts previously put forth to discuss changes in reaction times resulting from list composition were the de-emphasis account proposed by Monsell et al. (1992) and the time criterion account put forth by Lupker et al. (1997). The manipulation of foils in the Go/No-go

2 In support of this nding, a 2 2 2 ANOVA (Frequency Regularity Experiment) was run and

did produce a significant three-way interaction, F(1, 56) = 11.976, p = .001. However, given the specic

hypotheses of the current paper and to minimize Type I error rates that arise with multiple comparisons, we chose to constrain our analyses to the independent samples t test reported.

123

374 J Psycholinguist Res (2011) 40:367378

task provides evidence in favour of a de-emphasis of the sub-lexical system as proposed by Monsell et al. (1992). Participants in the Go/No-go task with pseudohomophones strategically decreased information processing along the sub-lexical system and relied primarily on lexical information. This was evidenced by the nding of a Frequency effect (i.e., resulting from the lexical system) but no Regularity effect (i.e., resulting from the sub-lexical system) and, subsequently, no Frequency Regularity interaction for the Go/No-go with

pseudohomophone task. In contrast, while the same strategy could have been benecial in the Go/No-go naming task with nonwords, the task did not necessitate such constraints to make a go/no go decision. Consequently, participants used information from both sub-lexical and lexical systems, as evidenced by the significant Frequency Regularity interaction.

Presenting stimuli in a manner that constrains response strategies demonstrates that skilled readers can focus primarily on lexical processing and initiate pronunciation prior to receiving information via the sub-lexical system.

Lupker et al. (1997) argued that the main problem with Monsell et al. (1992) ndings was that the effects were limited to high frequency words. However, our results showed a significant effect for low frequency exception words in the Go/No-go naming task with pseudohomophones. The time criterion account of Lupker et al. would be supported by a slowing of rapidly named stimuli and a speeding up of slowly named stimuli. However, while we found a speeding up of slowly named stimuli (e.g., low frequency exception words) in the Go/No-go naming with pseudohomophones, we did not nd a slowing of rapidly named stimuli in either experiment. In fact, the trend was for high frequency stimuli to be named faster in the Go/No-go naming task with pseudohomophones than in the Go/No-go naming task with nonwords (for exception words: 625ms vs. 667ms; for regular words 607ms vs. 672ms). Again, this supports the notion that the Go/No-go task with pseudohomophones encouraged a response strategy in which participants relied on information provided from the lexical system and proceeded onto speech without the information produced from the sub-lexical system. Overall, our results are more consistent with the de-emphasis account proposed by Monsell et al. (1992).

Our ndings support the notion that Frequency effects reect processing in the lexical system, whereas Regularity effects reect processing in the sub-lexical system. In the Go/No-go naming task with pseudohomophones, we eliminated the Frequency Regularity interaction

and any effect of regularity. These ndings are consistent with the interpretation that utilizing pseudohomophones in the Go/No-go task encouraged participants to rely heavily on lexical processing. Furthermore, the decreased reaction times for naming low frequency exception words in the Go/No-go naming task with pseudohomophones supports the notion that participants initiated overt responses prior to receiving the competing phonological codes produced from the sub-lexical system. These ndings provide converging evidence that participants strategically relied on lexical information.

Especially noteworthy is that the instructions and real word stimuli given to participants were identical in both experiments. All participants were instructed to name aloud only those stimuli that spelled a real word. While participants in the Go/No-go naming task with non-words could have relied solely on information from the lexical system, our ndings provide evidence that this was not the case. The significant Frequency Regularity interaction found

in the Go/No-go naming task with nonwords indicates that participants were utilizing information from both sub-lexical and lexical processing systems even though this approach is less efcient than relying solely on lexical processing. In contrast, the Go/No-go naming task with pseudohomophones, which forced participants to rely on primarily lexical information, produced significantly faster response times. Furthermore, our nding cannot be attributed to any differences that may exist between the regular and exception words utilized given that

123

J Psycholinguist Res (2011) 40:367378 375

the same stimuli were used in both experiments. That is, it was the nature of the foils that changed the nature of lexical and sub-lexical processing and not any characteristic of the regular or exception words. Together these results are important for researchers investigating task and strategy effects on resultant interactions between factors (e.g., see Braun et al. 2009; Ferguson et al. 2009; OMalley and Besner 2008; Stolz and Neely 1995 for discussions on control in the reading system using factors such as word frequency, stimulus quality and context). We provide direct evidence that control over the reading system is possible and while a particular strategy may be advantageous for improving performance, caution must be exercised when interpreting how participants are presumed to be formulating responses.

We provide the rst demonstration of an advantage for low frequency exception words when participants rely on information from the lexical route. Monsell et al. (1992) nding that low frequency exception words did not display an advantage when presented in pure blocks may be a result of participants incorporating some information from the sub-lexical system. While participants were instructed to rely on the lexical spelling of the stimulus, and the ratio of words to nonwords encouraged an emphasis of lexical information, the significant Frequency Regularity interaction found under these task conditions indicate that partic

ipants were still utilizing information from both routes. However, when participants must rely on lexical information only, because sub-lexical information will provide them with an incorrect pronunciation, we found that they adopted the appropriate strategy to complete the task.

Our results are consistent with the notion of strategic control via de-emphasis of information processing from the sub-lexical system. In keeping with Monsell et al. (1992) and dual-process models of reading (Borowsky and Besner 1993; Coltheart et al. 2001) we suggest that the lexical and sub-lexical systems continue to operate in parallel, though participants need not wait for the slowest output (i.e., from the alternative system, in this case, the sub-lexical system). This was shown by an elimination of the Regularity effect, which is hypothesized to represent sub-lexical processing. Second, we found an increase in the Frequency effect, which is hypothesized to represent lexical processing. The participants in our study were expected to rely solely on orthographic representations, and presumably lexical representations, as this strategy would be the most advantageous in both experiments. In addition, we further encouraged de-emphasis of the sub-lexical system by ensuring that the proportion of real words to be named was much greater than the proportion of nonwords to be ignored. However, participants in the Go/No-go naming task with nonwords still utilized the sub-lexical system, as evidenced by the Frequency Regularity interaction. In contrast,

when participants were forced to rely on lexical information, in the case of Go/No-go with pseudohomophones, we eliminated the Frequency Regularity interaction thereby provid

ing evidence that participants are reducing the involvement of the sub-lexical system. What is particularly noteworthy is the advantage for the low frequency exception words under these conditions. Consistent with the notion that participants were strategically relying on lexical information, reaction times in the Go/No-go naming task with pseudohomophones were faster overall, and there was a significant decrease in reaction times to name aloud the low frequency exception words.

We should note that while we are interpreting our ndings in terms of strategic reliance given the theoretical model proposed by Monsell et al. (1992), the notion of strategic reliance in reading remains controversial (Ferguson et al. 2009; Lupker et al. 1997; OMalley and Besner 2008). At present, the data presented here best ts with the interpretation provided by Monsell et al. (1992). Despite this, the current work provides valuable insight for furthering our understanding of basic reading processes and provides useful information regarding the contribution of lexical and sub-lexical information under varying list compositions.

123

376 J Psycholinguist Res (2011) 40:367378

It should also be noted that the nature of the Go/No-go task introduces a factor of inhibitory control, which may contribute to the differences reported in the present experiments. While the Go/No-go task has shown task dependent effects in imaging studies (see Simmonds et al. 2008), the resulting effects remain unknown in behavioural studies (Hino and Lupker 2000, 1998). Relevant to the present work, it is possible that the Go/No-go task with pseudohomophones and the Go/No-go task with nonwords differed in terms of difculty, whereby it was easier to suppress the response to pseudohomophones than nonwords. In this case, the inuence of inhibitory control processes and the subsequent change in difculty between the tasks should be reected in a constant reduction in reaction time. Yet, we also see the absence of an interaction in the Go/No-go with pseudohomophones suggesting that some factor beyond inhibitory control is inuencing reaction times for this task. The extent to which inhibitory control contributes to or modulates the change in response time between a Go/No-go task with nonwords and a Go/No-go task with pseudohomophones is a question for future research.

Conclusion

Our study provides a clear demonstration for the strategic separability of sub-lexical and lexical reading systems by showing: (1) Elimination of the Frequency Regularity interaction in

a Go/No-go naming task with pseudohomophones, and (2) De-emphasis of sub-lexical information due to inuence of the ller type, as evidenced by the elimination of the Regularity effect. These results are important to researchers investigating word recognition models as they provide critical information about Go/No-go naming tasks and the ability to manipulate sub-lexical information. Finally, our results also emphasize that while a particular strategy may be advantageous in a reading task, participants are not necessarily adopting that strategy.

Acknowledgment We would like to thank the Centre for Comparative Psycholinguistics at the University of Alberta for access and use of resources.

Appendix A

High frequency Low frequency

Regular words Regular words Nonwords Pseudohomophones dark brainfood bunch berv boarnfree coil boam braivgirl ditch boarm coaltgoes ame boke dawthad proud drose orehear sag dryn foartheat snatch fyce fynehome stack gerhn hawlfleave sweep ghyt gydemuch thrust hoaj hawttoo torn hoalt helledtwice truce hoert hoapwin mamths myndnyre pryd

123

J Psycholinguist Res (2011) 40:367378 377

Appendix A continued

Exception words Exception words ost stait does bought seafs stroal sofe swhis door breath theen terhn toov theem front broad tufe toon full dough vyfe truhmp give earn wawf tule gone learn woaf wyz nost sieveonce spreadone steakown sweatsays threadtouch tourtwo treadwhom

References

Balota, D. A., Yap, M. J., Cortese, M. J., Hutchison, K. A., Kessler, B., Loftis, B., et al. (2007). TheEnglish lexicon project. Behavior Research Methods, 39, 445459.

Besner, D., O-Malley, S., & & Robidoux, S. (2010). On the joint effects of stimulus quality, regularity and lexicality when reading aloud: New challenges. Journal of Experimental Psychology: Learning, Memory, and Cognition, 36, 750764.

Besner, D., & Smith, M. C. (1992). Models of visual word recognition: When obscuring the stimulus yields a clearer view. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 468482. Borowsky, R., & Besner, D. (1993). Visual word recognition: A multistage activation model. Journal of

Experimental Psychology: Learning, Memory, and Cognition, 19, 813840.

Braun, M., Hutzler, F., Ziegler, J. C., Dambacher, M., & Jacobs, A. M. (2009). Pseudohomophone effects provide evidence of early lexico-phonological processing in visual word recognition. Human Brain Mapping, 30, 19771989.

Coltheart, M., et al. (2001). DRC: A dual route cascaded model of visual word recognition and reading aloud. Psychological Review, 108, 204256.

Ferguson, R., Robidoux, S., & Besner, D. (2009). Reading aloud: Evidence for contextual control over lexical activation. Journal of Experimental Psychology: Human Perception and Performance, 35, 499507. Hino, Y., & Lupker, S. J. (2000). Effects of word frequency and spelling-to-sound regularity in naming with and without preceding lexical decision. Journal of Experimental Psychology: Human Perception and Performance, 26, 166183.

Hino, Y., & Lupker, S. J. (1998).The effects of word frequency for Japanese Kana and Kanji words in naming and lexical decision: Can the dual-route model save the lexical-selection account? Journal of Experimental Psychology: Human Perception and Performance, 24, 14311453.

Kinoshita, S., Lupker, S. J., & Rastle, K. (2004). Modulation of regularity and lexicality effects in reading aloud. Memory & Cognition, 32, 12551264.

Lupker, S. J., Brown, P., & Colombo, L. (1997).Strategic control in a naming task: Changing routes or changing deadlines? Journal of Experimental Psychology: Learning, Memory, and Cognition, 23, 570590.

Monsell, S., Patterson, K. E., Graham, A., Hughes, C. H., & Milroy, R. (1992). Lexical and sublexical translation of spelling to sound: Strategic anticipation of lexical status. Journal of Experimental Psychology: Learning, Memory, and Cognition, 18, 452467.

Olijnik, S., & Algina, J. (2003). Generalized eta and omega squared statistics: Measures of effect size for some common research designs. Psychological Methods, 8, 434447.

123

378 J Psycholinguist Res (2011) 40:367378

OMalley, S., & Besner, D. (2008). Reading aloud: Qualitative differences in the relation between stimulus quality and word frequency as a function of context. Journal of Experimental Psychology: Learning, Memory, and Cognition, 34, 14001411.

Paap, K. R., & Noel, R. W. (1991). Dual-route models of print to sound: Still a good horse race. PsychologicalResearch, 53, 1324.

Simmonds, D. J., Pekar, J. J., & Mostofsky, S. H. (2008). Meta-analysis of Go/No-go tasks demonstrating that fMRI activation associated with response inhibition is task-dependent. Neuropsychologia, 46, 224232. Stolz, J. A., & Neely, J. H. (1995). When target degradation does and does not enhance semantic context effects in word recognition. Journal of Experimental Psychology: Learning, Memory, and Cognition, 21, 596611.

Visser, T. A. W., & Besner, D. (2001). On the dominance of whole-word knowledge in reading aloud.Psychonomic Bulletin & Review, 8, 560567.

123