Running head: SECOND LANGUAGE PRODUCTION

Using the Native Language to Track Speech Planning in a Second Language

Rhonda McClain, Eleonora Rossi, & Judith F. Kroll

Author note
Rhonda McClain, Psychology Department, Pennsylvania State University; Eleonora
Rossi, Psychology Department, Pennsylvania State University; Judith F. Kroll, Psychology
Department, Pennsylvania State University.
Rhonda McClain is now at the Linguistics Department, Northwestern University;
Eleonora Rossi is now at Department of Psychology and Sociology, California State Polytechnic
University.
Correspondence concerning this article should be addressed to Rhonda McClain,
Department of Linguistics, Northwestern University, Evanston, Illinois. Phone: (847) 491-4533,
email: rhonda.mcclain@northwestern.edu.
This research was supported by: NSF Dissertation Improvement Grant 421-55-62910 to Rhonda
McClain, Judith Kroll, and Eleonora and NIH Grant HD053146 to Rhonda McClain. The writing
of this article was also supported in part by NSF Grants BCS-0955090 and OISE-0968369 to
Judith F. Kroll. The content is solely the responsibility of the authors and does not necessarily
represent the official views of the funding agencies. Rhonda McClain is now postdoctoral fellow
of Linguistics at Northwestern University, USA. The authors would like to thank Maya Waide,
Meaghan Wirtz, Nick Pollio, Christian-Berstein-Cuevos, and Manuel Pulido-Azpiroz for
research assistance.

Abstract

When proficient bilinguals prepare to speak the second language (L2), there is evidence for
inhibition of the more dominant language (L1). In the present study we asked how these
demands are manifest for L2 learners who are much more challenged at speaking the L2 (e.g.,
Costa & Santesteban, 2004). We examined production in English in two groups of native English
speakers who were learning Spanish as an L2. During a blocked language naming paradigm, one
group of learners named pictures in English (L1) after Spanish (L2) and the other named pictures
entirely in English (L1). We recorded EEG and behavioral responses while each group
performed the picture naming task. Like other recent studies, behavioral measures showed
inhibition of lexically-specific items (e.g., reduction in repetition priming). However, when
learners named in the L1 after the L2 the ERP record did not reveal an inhibitory effect. Learners
who spoke only the L1 experienced priming behaviorally and in the ERPs. The evidence
suggests that there are immediate consequences of speaking the L2 on subsequent L1 production
for learners, but no extended consequences when speech is planned in the L1 alone. Results are
consistent with an emerging view that bilinguals exploit multiple, distinct components of control
whose precise configuration is influenced by a combination of experiential and contextual
factors.

Keywords:

bilingualism, second language learning, language production, inhibitory control

Using the Native Language to Track Speech Planning in a Second Language

Introduction
Adult second language (L2) learners are notorious for being able to comprehend (e.g.,
(Midgley, Holcomb, & Grainger, 2009; Sunderman & Kroll, 2006), but not speak the L2 (e.g.,
Tokowicz, Michael, & Kroll, 2004). The handful of studies that have directly examined learners
production in L2 confirm that even producing single words in the L2 is a difficult task for
learners at early stages (e.g., Kroll, Michael, Tokowicz, & Dufour, 2002). One obvious source of
learners difficulty producing speech in the L2 is limited vocabulary. Another source of difficulty
is the large asymmetry in L1 and L2 dominance. This implies that even when learners know the
word they are attempting to speak in the L2, they experience interference from the L1. A
question of interest is what mechanisms exist that enable learners who are barely proficient at
speaking the L2 to produce words in face of competition from the dominant language.
A series of recent studies investigating speech in proficient bilinguals have provided
evidence suggesting that the L1 is momentarily suppressed, becoming slower and less automatic
when produced after immediately speaking the L2 (e.g., Branzi, Martin, Abutalebi, & Costa,
2014; Misra, Guo, Bobb, & Kroll, 2012; Van Assche, Duyck, & Gollan, 2013). One would not
expect even highly proficient bilinguals to experience interference in the L1, unless it were used
for the purpose of L2 regulation during speech planning. In the present study, we exploit the
insight gained by these studies and use the L1 as a window into the regulatory processes that are
exploited by L2 learners during L2 production. In the next section, we will revisit initial
evidence that shaped subsequent accounts of mechanisms available for L2 learners for regulating
language during speech production. Critically, although more recent studies differ from earlier
studies in precise implementations, there is converging evidence that in the context of speaking

the L2 the L1 undergoes change. In the present study, we extend this approach to examine the
mechanisms supporting L2 speech planning in L2 learners through the L1.
Evidence from Language Switching
Although few studies of speech production have focused on L2 learners who are
unequivocally beginners, some studies of language switching have made comparisons between
bilinguals who are still dominant in the L1 and bilinguals who are balanced in L1 and L2
proficiency. Costa and Santesteban (2004) compared the performance of L1 dominant bilinguals
who were still acquiring L2 proficiency and highly proficient bilinguals on the language
switching task. The researchers hypothesized that highly proficient bilinguals are able to select
the language they intend to speak early in speech planning without engaging inhibition, but L2
learners engage a great deal of inhibition in order to speak the L2. The pattern of results from
Costa and Santestebans (2004) study were consistent with their predictions. Less proficient
bilinguals demonstrated switch costs asymmetries. Highly proficient bilinguals who were
balanced in their language dominance revealed symmetric switch costs.
For many years, it was generally assumed that the differences in the pattern of switch
costs could be interpreted as evidence that inhibition is restricted to less proficient, unbalanced
bilinguals. In recent years, a number of criticisms have been raised about the interpretation of the
switch costs (Bobb & Wodniecka, 2013). In Costa and Santestebans (2004) study it is also clear
that all bilinguals experienced general slowing of the L1 under the imposed mixed language
conditions. One interpretation of the L1 slowing is that under conditions of frequent language
mixing, the L1 is suppressed for a long time to bias against its selection. The fact that the L1
becomes much slower after speaking the L2 contradicts the original interpretation by

demonstrating that all bilinguals experience inhibitory consequences when the L1 and L2 are
spoken in mixed language contexts.
It is important to recognize how strongly the claims made about the underlying
mechanism behind the switch cost asymmetry have been. Recently, more attention has been
given to the evidence that conflicts with the original interpretation. Critically, the presence of
what appears to be two different inhibitory effects in the same task suggests that language
switching tasks may be sensitive to multiple components of language control, with the switch
cost asymmetry indexing just one of those components. One additional limitation reflects the
generalizability of the task itself. Language switching always occurs in a mixed contexts.
Bilinguals do switch in naturalistic contexts, but not in the mixed fashion imposed by language
switching. Bilinguals are more likely switch languages after speaking one for an extended period
of time in naturalistic conversations. Compared with language switching tasks, tasks that mirror
the extended production in a single language would have the advantage of providing a controlled
experimental context to observe properties of language regulation while also exploiting more
natural conditions.
Evidence from Blocked Production
In a recent series of production studies, investigators have adopted precisely this
approach. These recent studies have used the blocked production paradigm to examine switching
when switches occur after an extended period of speaking a single language. The blocked picture
naming task affords the opportunity to test not only hypotheses about mechanisms of inhibition,
but also alternatives to inhibitory mechanisms (e.g., Philipp, Gade, & Koch, 2007; Philipp &
Koch, 2009). For example, one might hypothesize that under blocked language conditions
bilinguals do not show evidence of L1 inhibition since they can anticipate the language of

production. Misra et al. (2012) examined evidence for L1 regulation when bilinguals switched
languages during blocked picture naming, predicting that because the extended nature of the
production is a salient clue to the language that should be engaged bilinguals would not inhibit
the L1 after speaking the L2. In their study, they included a manipulation in which pictures were
repeated in the L1 and L2 blocks. Repetition of pictures typically produces priming such as faster
responses (e.g., Barry, Hirsh, Johnston, & Williams, 2001; Cave, 1997; Cave & Squire, 1992;
Mitchell & Brown, 1988) and evidence of neural facilitation. Therefore, the critical prediction
was that they would observe priming of repetitions when the L1 followed the L2.
In Misra et al.s (2012) study, two groups of bilinguals performed blocked picture
naming while event-related potentials (ERPs) were recorded. One group named in L1 before L2
and the other named in L1 after the L2. To determine how long regulation lasted, the language
switch included two continuous blocks of production. ERPs revealed two different patterns,
depending on whether the L1 was spoken before or after L2. Bilinguals who spoke the L2 after
the L1 demonstrated evidence for priming (reduced negativity in N200 and more positivity in
LPC components). Bilinguals who named in the L1 after the L2 showed increased negativity in
the N200 that extended into the later ERP window, suggesting inhibition.
The results of Misra et al.s study demonstrate that speaking the L2 before the L1
eliminates evidence for priming. The absence of priming was surprising at the time given the
assumption that speaking a single language for a sustained period of time would have set up a
context where inhibition is not needed. This pattern supports the idea that the L1 is generally
suppressed in order to speak the L2. The extended time course of the effects also showed that,
even for proficient bilinguals, speaking the L2 is not a matter of momentarily suppressing the L1.
However, the design of Misra et al. could not provide information about the scope of language

regulation. Language regulation could affect only a specific set of words spoken, words that
share the same semantic network, or the whole language itself.
In a recent blocked production study, Van Assche et al. (2013) introduced modifications
that enabled them to discriminate processes of language regulation at different levels of scope. If
the scope of regulation were lexically-specific, they expected only specific words that have been
spoken in the context to be affected by language switching. If the scope of regulation generalized
to the entire language, even items that had never been produced would be affected by language
switching. Bilinguals performed verbal fluency in sets blocked by the language of production.
Like Misra et al.s study, some bilinguals were assigned to speak the L1 before the L2 and some
produced the L1 after the L2. To assess the scope of regulation, verbal fluency categories were
either repeated across the language blocks or completely new. Here, repetitions were used to
assess whether priming had been affected by the requirement to switch languages. Evidence for
whole language and lexically-specific regulation was observed. One group of bilinguals showed
only evidence of lexically-specific inhibition, whereas another group of bilinguals showed
evidence for inhibitory effects at both levels of scope. This result suggests that inhibitory
mechanisms dissociated.
At present it is unclear whether there are experiential factors that drive different patterns
at each level of scope and how to characterize these factors. What is interesting about recent
studies is that, although they provide evidence that largely favors inhibition, they also suggest
that there is not a single mechanism of inhibition. Mechanisms of language regulation appear to
vary in the scope and time course and not all consequences of switching languages appear to be
inhibitory. Therefore, these findings lead to the prediction that bilinguals have access to a much
more diverse set of mechanisms than those described in existing accounts of bilingual language

control. We can already anticipate that the mechanisms of regulation would have to be strong
because learners are so dominant in their native language. In light of more recent findings, the
issue at hand is not whether inhibition is present or absent when learners plan to speak L2, but
how control manifests with respect to the precise scope and time course properties. Studying
consequences of speaking the L2 for the L1 in learners might provide clearer evidence then on
the different types of mechanisms since control should reverberate more strongly on the L1.
In the present study, we took a new approach to investigating L2 planning processes in
L2 learners by focusing not on the L2 directly. We hypothesized that the mechanisms that were
engaged for enabling production in L2 will spillover to affect subsequent L1 production and
serve as the window into mechanisms of L2 production themselves. Two ERP experiments were
used investigate the behavioral and neurophysiological signatures of L2 speech planning. Two
groups of L2 learners of Spanish were tested. One group of learners performed the blocked
picture naming task in the L1 after L2 order and another group of learners named pictures only in
the L1. Like Misra et al. (2012), we examined the extended time course of the consequences of
speaking the L2 for the L1. Similarly to Van Assche et al. (2013) we included manipulations to
examine the scope of language regulation. We asked if switching languages affected specific
words or generalized to even new words. We also combined the blocked picture naming
paradigm with the event-related potential technique to examine fine-grained changes in
processes of speech planning that occur in this task.
Experiment 1
In Experiment 1, we investigate the consequences of speaking the L1 after the L2 using a
modified blocked picture naming paradigm. L2 learners first named a block of pictures in L1 and
a second block of pictures in L2, followed by six subsequent blocks of naming pictures in L1

only. The addition of six L1 naming blocks after production in the L2 permits a more finegrained analysis of the time course language regulation and the scope of the consequences of
speaking the L2 on L1. The six blocks which followed the intervening L2 block included
pictures that were previously named in either the L1 or the L2 blocks (i.e., cat, pan) and
completely new pictures. Thus, the present study mirrors the design of previous studies that
demonstrated lexically-specific and whole-language patterns of inhibition in response to naming
identical repetitions and new items respectively (e.g., Branzi et al., 2014; Van Assche et al.,
2013).
Like previous studies, we exploited the fact that repetitions typically produce priming to
infer whether the nature of regulation was inhibitory. Generally, repeating pictures facilitates
subsequent naming (e.g., Barry, Hirsh, Johnston, & Williams, 2001; Cave, 1997; Cave & Squire,
1992; Mitchell & Brown, 1988). We can infer that the L1 has been inhibited if speaking the L2
reduces or eliminates the repetition benefit. Alternatively, priming may persist even when the L1
is spoken after the L2. If this were so, responses would become facilitated at the behavioral level
and neural level. To index the presence of facilitation at the neural level, we focus on the N400
component. This component has frequently been cited in studies manipulating semantic or
stimulus-level repetition priming (e.g., Franklin, Dien, Neely, Huber, & Waterson, 2007;
Henson, Rylands, Ross, Vuilleurmeir, & Rugg, 2004; Kutas & Hillyard, 1989; McPherson &
Holcomb, 1999). Neurally, priming should manifest as a typical N400, with less negativity in the
waveforms emerging for identical repetitions relative to new pictures. If there is inhibition, the
direction will be the opposite. Identical repetitions will elicit greater negativity relative to new
pictures. To index inhibition, we focused on the N200 component. The N200 is a negative going
peak in the waveform, which generally in cognitive tasks, appears between 250 and 350 ms after

the onset of the stimulus and is usually frontally distributed on the scalp. It has been
hypothesized to reflect response conflict or inhibition (e.g., Falkenstein, Hoormann, &
Hohnsbein, 1999; Nieuwenhuis, Yeung, van den Wildenberg, & Ridderinkhof, 2003) because it
has been found in tasks that require response suppression (such as Go/No-go tasks). An
additional motivation for focusing on the N200 is that it is typically the component of interest in
studies examining language regulation (e.g., Christoffels et al., 2007; Jackson et al., 2001,
Verhoef, Roelofs, & Chwilla, 2009). For example, Misra et al. 2012 identified an early N200, in
the time window of 200 to 375 ms, that was modulated by the requirement to speak the L1 after
the L2. Therefore, the window we selected was consistent with the time course of the N200
reported in prior bilingual language production studies.
As mentioned previously, it is possible that consequences of speaking the L1 after
speaking the L2 may be distinct at each level of scope. If speaking the L2 affects the entire L1,
the L1 would experience changes in sensitivity to priming even when completely new pictures
were named (i.e., truck). A more selective alternative is that the consequences of speaking
the L1 following the L2 are lexically-specific, affecting the ability to produce specific words in
L1 that were previously named in L2 (i.e., cat). We also borrowed the logic of previous studies
to track the extended time course of language regulation. If the L1 were affected by speaking the
L2, then it would be possible to determine precisely how long the L1 was affected by examining
behavioral and neural patterns over each of the later L1 blocks.
Method
Participants
Eighteen students of Penn State University participated in Experiment 1. All participants
were right-handed, had normal or corrected-to-normal vision, and no history of neurological

disorder. Participants were native English speakers with intermediate-level proficiency in

Spanish. They were all late learners who had studied Spanish since age 12 on average and had a
minimum of two college semesters of Spanish instruction. Table 1 summarizes the language and
cognitive profile of the participants who were tested in Experiments 1 and 2.

-------------------------------Insert Table 1 about here

--------------------------------

Materials
One hundred and ninety-two pictures were selected from nine semantic categories:
Animals, Body Parts, Clothing, Foods, Furniture, Kitchen Items, Tools, Vegetables, and
Vehicles. Each picture was displayed in line drawing, color photograph or black and white
photograph format in order to reduce the priming associated with presenting an identical picture.
The colored and black and white photograph images were retrieved from Google Images. Line
drawings were taken from the Snodgrass and Vanderwart (Snodgrass & Vanderwart, 1980)
picture database and Google Images. Of the total picture stimuli, approximately 40 percent were
selected from Google Images and had been normed in a previous study (Rossi & Kroll, in
preparation). Pictures were controlled for their dimensions, and were 400 x 400 pixels. A set of
486 picture stimuli was created.
There were a total of eight blocks presented during the task. Pictures were assigned to
conditions that targeted the local and global levels of language control. Pictures were matched as
closely as possible across conditions for the number of syllables in the pictures name,
frequency, familiarity, and number of taxonomic features, which are variables that have been
found to influence picture naming latencies in past research (e.g., Alario et al., 2004; Cuetos,

Ellis, & Alvarez, 1999; Snodgrass & Yuditsky, 1996). The characteristics of the pictures are
summarized in Table 2.
-------------------------------Insert Table 2 about here
--------------------------------

Pictures were presented during eight blocks (Block1-8). Twenty-seven pictures were
selected for presentation during Block 1. Then, 27 additional pictures were selected for
presentation during Block 2. Pictures that first appeared during Block 1 or Block 2 were named
again during Blocks 3-8 as lexically-identical repetitions from Block 1 and lexically-identical
repetitions from Block 2. In addition to lexically-identical pictures, a set of entirely new pictures
drawn from categories not previously seen were included in the later blocks of the experiment
(new). To ensure that lists were matched properly, all pictures were assigned to condition lists
using a matching algorithm (Match.exe, Van Casteren & Davis, 2007).
Pictures were assigned to lists, rotating when they appeared during the eight blocks of the
experiment. In addition, the later L1 blocks (L1 after L2) were counterbalanced across
participants to ensure that items were named in different orders. To adequately match the lexical
properties of the pictures that appeared in the six later test blocks, these later blocks were divided
into three blocks (e.g., Blocks 3 and 4, Blocks 5 and 6, and Blocks 7 and 8). Three pairs of
blocks were retained to ensure that it would still be possible to detect graded changes to the L1
after speaking the L2. During the later blocks (Blocks 3-8) pictures were cycled so that a picture
only appeared within one of the three later test blocks. For example, if shoe was an item
assigned to Blocks 3 and 4 as a lexically-identical repetition, the picture was named first in

Block 1, then named during Blocks 3 and 4, but then could not be presented as a stimulus later in
the experiment.
Procedure
Participants were tested individually in a sound-proof testing room. They were seated in a
comfortable chair approximately 20 inches away from a Dell computer monitor. A microphone
attached to the S-R box was placed directly in front of the participant as closely as possible
without occluding their vision. Participants first received instruction in English on the computer
screen. They were informed that a series of pictures would appear on screen and that they would
name pictures during some sections in the L1 and some sections in the L2. At the beginning of
each trial, a fixation sign (+) was presented at the center of the computer screen for 500 ms. After
the offset of the fixation sign, a picture was presented. The picture was presented for 1000 ms or
until the participants responded. After they responded, a blank screen was presented for 1000 ms
and a fixation sign appeared again. If participants did not know the name of the picture, they
were instructed to be silent. Participants received 10 practice trials of English picture naming
before Block 1 and 10 practice trials of Spanish naming before Block 2. There was no
familiarization with picture names prior to testing.
ERP Recording and Data Analysis.
The continuous electroencephalogram (EEG) was recorded using a 32-channel sintered
Ag/AgCl electrode array mounted in an elastic cap according to the 10-20 system (QuikCap,
Neuroscan Inc.). Recordings were referenced to the left mastoid during recording and rereferenced off-line to the average activity of the left and right mastoids. Electrode impedances
were kept below 5 k. Lateral eye movements were measured by electrodes placed on the outer
canthus of each eye. Vertical eye movements were measured by electrodes placed on the upper

and lower orbital ridge of the left eye. Eye recordings were later used off-line to reject
contaminated trials. The electrophysiological signals were amplified using Neuroscan Synamps
with a band pass filter of 0.05 to 100 Hz and a sampling rate of 500 Hz. Only trials with correct
responses were included in the analyses. A pre-stimulus baseline of 150 ms and an epoch
duration of 600 ms post-stimulus were used to compute average ERPs per condition. Trials with
eye movement artifacts or blinks and peak-to-peak deflections over 100 V were rejected. A
digital low-pass filter of 30 Hz (24 dB/oct) was applied when analyzing the data off-line.
Based on visual inspection of the waveform and the previous literature (Strijkers, Costa
& Thierry, 2009; Christoffels, Firk, & Schiller, 2007; Misra et al., 2012) two time windows were
selected for further investigation. Mean amplitudes were calculated over each of the time
windows corresponding to the component of interest, an early negativity, or N200, between 200
and 270 ms, and the N400, between 350 and 450 ms.
Using the criteria described, only correct, artifact-free trials were included in the ERP
analyses. To allow for adequate signal-to-noise ratio, the minimum number of trials for
computing average ERPs for each condition in each participant was 15. For the participants
included in the final sample, 15.17% of trials were rejected after excluding trials with errors and
artifacts (8 %, 3%, 11%, 10%, and 12 % for Block 1, Block 2, Blocks 3 and 4, Block 5 and 6,
and Block 7 and 8 respectively and 6%, 10%, 13%, 9%, and 13 % for new, new from Block 1,
new from Block 2, repeated from Block 1, and repeated from Block 2 respectively). For each
component, mean amplitudes over set time windows were computed as dependent variables. For
the first set of analyses, for each dependent component, repeated measures ANOVAs comparing
language (L1 or L2) were computed over separate electrode groupings.

Based on the topographical distribution of the effects we were interested in, separate
analyses were completed for the midline and lateral sites. For the midline sites, four levels of
electrode site (FZ, CZ, PZ, and OZ) were included as a second variable. For the medial-lateral
sites, variables of hemisphere (2: left/right) and electrode site (5 levels: FP1/FP2, F3/F4, C3/C4,
P3/P4, O1/O2) were factored into the ANOVA. For the lateral-lateral sites, variables of
hemisphere (2: left/right) and electrode site (3 levels: F7/F8, T7/T8, P7/P8) were factored into
the ANOVA. The GreenhouseGeisser correction was applied to account for non-sphericity of
the data (Greenhouse & Geisser, 1959); uncorrected degrees of freedom and corrected
probabilities are reported. Only results including the factors of language, block, category
repetition, or lexical repetition as main effects or interactions between these factors and electrode
site and/or hemisphere are reported.
Behavioral Data Analysis
Recorded picture naming responses were transcribed and scored for accuracy. Responses
that included the expected name of the picture, synonyms, or close semantic relatives were
counted as correct. Responses that were unrelated to the expected picture name, responses that
started with an article or hesitation, and omissions were scored as errors. Responses that the
microphone did not detect were eliminated as technical errors. After trimming response latencies
for correct responses that were less than 300 ms or greater than 3000 ms for both L1 and L2
trials, response latencies that were 2.5 standard deviations above or below the mean were
identified as outliers and excluded from the analyses. Finally, accuracy and response latencies
for correct responses were calculated. Errors (8.4 %), outliers (4.4 %), and technical errors (0.91
%) were excluded from the analyses.

Results
In what follows, we report the behavioral and ERP results to address our predictions. For
all the analyses we will first report the behavioral results and then the ERP results.
Analysis 1. Behavioral and neurophysiological correlates of L1 and L2 naming: Block 1 vs.
Block 2
Blocks 1 and 2 establish precisely how much learners who are unskilled at speaking the
L2 differ in the effort required to speak L1 and L2. Learners who are still in the process of
acquiring the L2 were expected to be slower and less accurate naming pictures in the L2 than in
the L1. At the neural level, we expected the comparison of L1 and L2 naming during Blocks 1
and 1 to reveal early divergences in the ERP waveforms that extended to the later time windows
(e.g., Misra et al., 2012; Hoshino & Thierry, 2011). Furthermore, we expected that the magnitude
of these differences in ERPs would be especially large for learners, surpassing the effects
observed when L1 and L2 production is compared for proficient bilinguals (e.g., Strijkers et al.,
2009).
Behavioral results: A one-way ANOVA was performed on accuracy and response
latencies with block as a within subject factor (Block 1; Block 2). Results show a significant
main effect of block for accuracy, and for naming latencies, revealing that L2 learners were
significantly less accurate and slower when they named pictures in the L2.

-------------------------------Insert Figure 1 about here

--------------------------------

ERP results: N200 component. At the midline, the main effect of language did not reach
significance, although there was a trend for an interaction between block and electrode, which
revealed that the N200 component was less negative in Block 2 (L2 naming), than in Block 1 at
posterior electrodes. At lateral-lateral sites a significant two-way interaction between block and
electrode indicated that the N200 was less negative in Block 2 at the parietal electrodes.
N400 component: There were no significant main effects or interactions.
In general, differences in the neural processes underlying speech production the L1 and
L2 were modest compared to patterns observed for highly proficient bilinguals (e.g., Strjikers et
al., 2009). Because learners made many omissions, it is possible that learners did not sufficiently
activate the L2. If later effects of speaking the L2 on the L1 depend on successful naming in the
L2, then learners might show limited consequences of speaking the L1 after the L2. However, if
the consequences for the L1 are driven by the attempt to speak the L2, then the intention to name
a picture in Spanish should prompt regulation of the L1 (e.g., Green, 1998; Strijkers, Holcomb &
Costa, 2011).
---------------------------------------------Insert Table 3
----------------------------------------------

Analysis 2. Testing the lexical-specific scope of inhibition: comparing new items (Blocks 38) vs. lexically-identical repetitions from Blocks 1 and 2 (Blocks 3-8)
Behavioral results: To address the time course of regulation and to ask whether learners
recovered the L1, we examined the magnitude of priming at each of the later blocks. The
ANOVA performed on the accuracy showed a marginally significant main effect of lexical
repetition, indicating that participants were facilitated when the named naming repetitions from
Block 1. They were not facilitated when they named repetitions from Block 2. For response

latencies, there was a significant main effect of lexical repetition. Further paired-sample t-tests
indicated that participants named repetitions from Block 1 faster than new pictures. Participants
did not name repetitions from Block 2 faster than new pictures. The strikingly different patterns
of priming observed for suggest that the learners speech was affected by a mechanism
suppressing lexically-specific representations.

---------------------------------------------Insert Figure 2
------------------------------------------------------------------------------------------Insert Table 4 about here
----------------------------------------------

Results from response latencies revealed an interaction between block and lexical
repetition that was reliable by-items, but not by-subjects. Repetitions from Block 2 were named
slower than new pictures in blocks 3-4, and blocks 5-6. However, during the last L1 block
(Blocks 7 and 8) repetitions from Block 2 produced facilitation. The results converge with
previous evidence, demonstrating that the mechanism of lexically-specific inhibition is very
long-lasting.
ERP results: N200 component. At the midline, there was a significant main effect of
lexical repetition that was qualified by a significant two-way interaction between lexical
repetition and electrode. Follow-up ANOVAs conducted at each electrode site showed that
repetitions from Block 1 and repetitions from Block 2 elicited less negativity in the waveforms
than new items at each of the electrodes. There were no differences between repetitions from

Block 1 and Block 2 at any of the electrode locations. There were no other significant main
effects or interactions.
N400 component. At the midline, there was a marginally significant main effect of block
and a significant main effect of lexical repetition. Furthermore, there was a significant interaction
between lexical repetition and electrode. ANOVAs performed over each electrode indicated that
repetitions from Block 1 and Block 2 elicited less negativity in the waveform than new pictures
at central, parietal and occipital electrodes. An identical pattern was observed at the mediallateral electrodes. There was no interaction between lexical repetition and block, suggesting that
the priming was consistent throughout the task.
---------------------------------------------Insert Table 5 about here
----------------------------------------------

Analysis 3. New items at baseline (Block 1) vs. New items in the later blocks (Blocks 3-8)
Behavioral results: A one-way ANOVA on accuracies and response latencies with four
levels of block (Block 1, Blocks 3 and 4, Blocks 5 and 6, Blocks 7 and 8) for accuracy and
response latency was performed. For the accuracy, there were no significant results. However,
the analysis of response latencies revealed a significant main effect of block. In the analysis by
subjects, new pictures were named more slowly in each of the later L1 blocks than in Block 1.
One interpretation of this pattern is that speaking the L2 prompted inhibition of the entire L1.
However, we will reserve further commentary on the interpretation of these results until the
General Discussion.
---------------------------------------------Insert Figure 3
----------------------------------------------

ERP results: ANOVAs were performed to compare the mean amplitudes of the waveform during
Block 1 to completely new pictures that were named during Blocks 3 and 4, Blocks 5 and 6,
Blocks 7 and 8.
N200 component. At the midline, there were no significant main effects or interactions.
Marginally significant main effects of block and interactions between block and electrode were
observed at the medial-lateral electrodes and lateral-lateral electrodes. These results
demonstrated that the later blocks elicited increasing negativity as the task progressed. Further
inspection of the interactions revealed that this effect was largest at the frontal electrodes.
N400 component. At midline and medial-lateral sites, there were no significant main
effects or interactions. At the lateral-lateral sites, there was a significant main effect of block that
was qualified by a three-way interaction between block, electrode and hemisphere. Follow-up
ANOVAs conducted over each electrode site indicated a marginally significant interaction
between block and hemisphere at temporal electrodes. New pictures generated less negativity in
the later blocks relative to Block 1. The reduction in negativity was largest in the left
hemisphere, suggesting a more left lateralized response.
---------------------------------------------Insert Table 6
----------------------------------------------

Interim discussion: Experiment 1

In Experiment 1, we sought evidence for the mechanism that enables learners to speak
the L2 by examining the consequences of speaking the L2 for the L1. We also evaluated the
claim made by some models of bilingual production which states that regulation by inhibition
would be even more critical for L2 learners than for proficient bilinguals. Overall, we observed

extended consequences of speaking the L2 on subsequent L1 picture naming that reflected both
inhibition and priming. Behavioral measures revealed a mixture of priming and inhibitory
effects. Learners showed priming when they named repetitions of pictures that were previously
named in the L1. However in other conditions, facilitation was eliminated where priming was
expected. Repetitions of pictures that were previously named in L2 (Spanish) did not show
evidence of priming until the very last block. This pattern is consistent with the presence of a
lexically-specific inhibitory mechanism.
At the behavioral level, there was clear evidence for inhibition of pictures that were
previously named in the L2. However, the ERPs during both the N200 and N400 time windows
revealed less negativity in the waveforms when repeated pictures from Block 2 were named,
suggesting facilitation. One possibility is that the diverging patterns in ERPs and behavioral
measures reflects the learners limited ability to overtly name the L2 pictures. We will return to
this idea in the General Discussion. In sum, the results of Experiment 1 suggest, despite low
accuracy in speaking L2, learners were affected by the immediate requirement to speak the L2.
The observation that the requirement to speak the L2 would not only eliminate, but reverse
priming for repetitions is surprising. Such findings are difficult to accommodate under the
assumption that subsequent speech in the L1 was not influenced by the attempt to speak the L2.

Experiment 2: L2 learners speaking only the L1

In Experiment 2, a separate group of L2 Spanish learners performed the same blocked
picture naming task, but spoke only the L1 during the task. This enabled us to determine whether
or not the behavioral patterns of inhibition observed at the level of lexical or whole language
representations levels were a consequence of the requirement to speak the L2. We hypothesized

that at this relatively early stage of bilingualism, learners should not be affected by the L2 when
they are not required to actively engage it. Therefore, learners speaking only the L1 should not
show neural and behavioral reduction in priming.

Method
Participants were 19 learners of L2 Spanish who attended Penn State University at the
time of the study. L2 learners were all native speakers of English who had acquired intermediatelevel knowledge of Spanish. Participants were recruited so that they would match learners from
Experiment 1 on a number of characteristics. Table 4 displays the linguistic and cognitive
characteristics of the learners who participated in Experiments 1 and 2. There were no
differences in age (t(35)=-.05, p=.96) nor level of education (t(35)=.10, p=.92) between groups.
Self-rating of L1 proficiency and L2 proficiency did not differ between groups (t(35)=-.97,
p=.34 and t(35)=1.15, p=.26, respectively).
Materials
The materials were identical to those used in Experiment 1.
Procedure
The procedure was nearly identical to Experiment 1, with the exception that picture
naming was performed entirely in English (L1).
Behavioral data analysis.
For the data analyses, behavioral results were cleaned using the same set of criteria as
Experiment 1. Errors (4 %), outliers (2 %), and technical errors (2%) were excluded from the
data of L2 learners speaking the L1 alone.
ERP data analysis.

ERP data analysis followed the procedure of Experiment 1.

Results

Analysis 1. Examining the L1 at baseline: Block 1 vs. Block 2

Behavioral analysis: In the by-subjects analysis, the main effect of block was not
significant in response latencies. There was a significant effect of block in the item analysis
showing that participants named pictures more accurately in Block 1 than Block 2.
---------------------------------------------Insert Table 7 about here
----------------------------------------------

ERP analysis: N200 component. The main effect of block did not reach significance at
any of the electrode columns. Nor did the interaction between block and electrode. However,
there was a significant two-way interaction between block and hemisphere at the medial-lateral
and lateral-lateral columns. At the medial-lateral columns, increased negativity was found in the
waveforms for Block 1 relative to Block 2. This analysis also revealed an interaction with
hemisphere, such that significant amplitude differences between Block 1 and 2 were present only
in the left hemisphere. At the lateral-lateral sites, a three-way interaction between block,
hemisphere, and electrode revealing a different pattern. At the lateral-lateral electrodes, Block 1
elicited less negativity than Block 2.
N400 component. At the lateral-lateral sites, there was a marginally significant two-way
interaction between block and hemisphere indicating and less negativity in the right hemisphere
for Block 1 relative to Block 2.
---------------------------------------------Insert Table 8 about here

----------------------------------------------

We expected that there would be little, if any, evidence of differences during Block 1 and
Block 2 for L2 learners naming in the L1 alone. Instead, the behavioral and ERP results
demonstrated significant differences during Block 1 and Block 2 naming. Because Block 1 was
named more accurately, it is possible that the ERP differences similar reflect more efficient
naming during Block 1 than Block 2.
Analysis 2. Examining the effect of lexical repetition: L1 in the later blocks (Blocks 3-8)
Behavioral analysis: The ANOVA performed on response latencies revealed a significant
main effect of lexical repetition. Further paired-sample t-tests indicated that participants named
repetitions from Block 1 and repetitions from Block 2 faster than new pictures. As predicted,
there was no difference in response latencies between repetitions from Block 1 and Block 2.
In the accuracy analysis, the main effect of lexical repetition was significant by subjects,
and significant by items. Consistent with the analysis of latencies, there was no difference in the
amount of priming produced by repetitions from Block 1 and Block 2.
---------------------------------------------Insert Figure 4 about here
----------------------------------------------

ERP analysis: N200 component. At the midline and lateral-lateral sites, there was a significant
main effect of lexical repetition. Both types of repetitions elicited less negativity relative to new
items and the two types of repetitions did not differ in the amount of priming. At the mediallateral electrodes, there was no main effect of lexical repetition, but a significant interaction
between block and lexical repetition. The interaction showed that priming was significant for
both types of repetitions, but priming peaked during different blocks for the two type of
repetitions.

N400 component. At the midline, there was a marginally significant main effect of block.
At the lateral-lateral sites there was a marginally significant four-way interaction between block,
electrode, hemisphere and lexical repetition. Most relevant to the predictions, the interaction
revealed that during the later blocks, repetitions from Block 1 and Block 2 tended to elicit less
negativity in the waveform than new pictures. Interestingly, the reduction in the negativity of the
waveform for repetitions was modest during Blocks 3 and 4. Priming increased during the
remaining blocks.
---------------------------------------------Insert Table 9 about here
----------------------------------------------

Analysis 3. New items at baseline (Block 1) vs. New items in the later blocks (Blocks 3-8)
Behavioral analysis: A one-way ANOVA on response latencies examining the effect of
block revealed a marginally significant main effect of block in the analysis by subjects. There
was a trend indicating that new pictures were named more slowly in each of the later blocks than
when pictures were named during Block 1. In the analysis of accuracy, the main effect of block
was significant. Post-hoc paired-sample t-tests indicated that participants named new pictures
during Blocks 3 and 4 and Blocks 5 and 6 less accurately than during Block 1.
Contrary to predictions, behavioral patterns did not indicate evidence that was across the
board consistent with priming. The ability to name new pictures appeared to suffer in the later
blocks. In the later blocks, learners were significantly slower and less accurate than when they
named pictures in Block 1. One explanation for this pattern could be that the mixing of
completely new and old pictures could have influenced cognitive control and memory retrieval
processes. Block 1 and the later blocks differ in degree of mixing, with the later blocks being
made up of a large number of semantic categories. Elsewhere in the memory literature, this

mixing effect has been likened to setting a special retrieval mode that imposes an additional
step of checking whether an item is completely new (e.g., Ecker & Zimmer, 2009). This added
step may account for the longer response latencies during the later blocks, though the behavioral
data alone do not offer clarification.
-------------------------------Insert Figure 5 about here
--------------------------------

ERP analysis: N200 component. There was a significant main effect of block at the
midline and a marginally significant effect at the medial-lateral columns. The main effect of
block indicated that each of the later L1 blocks elicited greater negativity in the waveform than
Block 1. At the medial-lateral columns, there was also a marginally significant interaction
between block and electrode that was maximal at the posterior electrodes. At the lateral-lateral
sites the interaction between block and hemisphere reached significance. In the right hemisphere,
the difference between Blocks 1 and the later blocks was maximal during Blocks 3 and 4. In the
left hemisphere, the difference was largest during Blocks 5 and 6.
N400 component. At the lateral-lateral sites, there was a significant interaction between
block and hemisphere and a marginally significant three-way interaction between block,
hemisphere, and electrode. The interaction revealed that the difference between Block 1 and the
later blocks was greatest during Blocks 5 and 6. However, increased negativity in the later blocks
was more pronounced in the left hemisphere than the right and over the frontal and temporal
electrodes than the posterior electrodes.
---------------------------------------------Insert Table 10 about here
----------------------------------------------

Discussion
Experiment 2 was conducted to interpret the underlying basis of the effects observed in
Experiment 1. A separate group of L2 learners performed the same extending blocked picture
naming task, but under the requirement to speak the L1 alone. One prediction was that learners
who spoke only the L1 would demonstrate priming from literal repetitions. We predicted that
there would be no evidence of a global slowing to the L1, but that the L1 would appear
unchanged across the initial and later blocks.
When learners produced the L1 alone, we observed robust priming for literal repetitions
at the behavioral and neurophysiological level. Because Experiments 1 and 2 differed only in the
requirement to speak L2, one can infer that learners did indeed engage inhibition when speaking
the L2. The results suggest that the reversal of priming for lexical repetitions when learners
spoke the L1 after the L2 reflected inhibition.
However, the results of Experiment 2 suggest that the slowing of the L1 does not reflect
inhibition. When L2 learners produced in the L1 alone, the L1 was named more slowly in later
blocks. In addition, there was reduced negativity in the waveform for the later L1 blocks relative
to the initial L1 block. These findings suggest that there other mechanisms could underlie the L1
slowing. We speculate that the mixing of new and repeated items in the later L1 blocks promoted
a distinct set of memory and cognitive control processes than those that occurred during Blocks 1
and 2.
General Discussion
Few studies have examined the mechanisms that enable L2 speech production in learners
who are still acquiring an L2. For the first time, we demonstrated that mechanisms of L2

production can be identified by examining consequences of attempting to speak the L2 for the
L1. In two experiments, less proficient English learners of Spanish completed a blocked picture
naming task. ERP and behavioural measures were used to identify the mechanisms that enable
L2 learners to speak the L2. The evidence from the current study is pretty clear. Changes to the
speed and accuracy of L1 production and reduction of priming occurs when the L1 is preceded
by even a brief period of L2 production. Therefore, the blocked picture naming paradigm
provided robust and reliable evidence for regulation of the L1 and demonstrated that it is
possible to study L2 production even when L2 production abilities are relatively poor. The
results demonstrate the power and utility of the blocked picture naming paradigm.
We will now summarize the patterns of results found when learners spoke the L1 after the
L2 and the L1 alone. We begin by evaluating whether the results force us to reject a model that
predicts greater inhibition for L2 learners, such as Costa and Santestebans (2004) account of L2
speech planning. We then construct an account of how language regulation develops with respect
to mechanisms of cross-language interaction. Before concluding, we consider the separate
contribution of task factors to changes within the production system. Finally, we propose future
research directions.
The role of inhibitory processes in L2 learners speech planning
In Experiment 1, the main prediction was that the requirement to speak the L1 after the
L2 would prompt L1 inhibition. Costa and Santesteban (2004) argued that L1 inhibition is
strongest for L2 learners and less proficient bilinguals. This initial interpretation of these data
had a lasting impact on theories of bilingual language control, even though some views have
been revised (Costa, Santesteban, & Ivanova, 2006). The conventional view has been that
learners recruit inhibitory processes more strongly than proficient bilinguals. We revisited Costa

and Santestebans claim, but this time examining the consequences of speaking the L2 for
subsequent L1 production using the blocked picture naming task.
When L2 learners named pictures in the L1 after the L2, they showed evidence for
inhibition of lexically-specific items. To index lexically-specific inhibition, we repeated pictures
that were previously named in the initial blocks of L1 or L2 production. The clear prediction is
that presentation of a literal repetition should produce priming, or facilitation in naming at the
behavioral and neural levels. We did not observe behavioral priming in the later L1 blocks for
literal repetitions of pictures initially named in the L2. Behaviorally, the absence of priming
persisted across four out of the six individual blocks. This implies that a strong inhibitory process
must have countered any facilitation from repetition. To index inhibition for the whole language,
we compared the ability to name entirely new pictures during each of the later blocks (after the
L2 was produced) to naming pictures during the initial block of L1 production (before the L2
was produced). Learners showed slower response latencies in all of the L1 blocks of production
that followed production in the L2. However, additional evidence from other conditions within
the present study suggested that the slowing of the L1 after L2 production did not reflect
inhibition. For example, the ERP record indicated a slight reduction in negativity during the
N400 time window for the blocks of L1 production that followed L2 production. This suggests
that, when learners plan speech, inhibition does not generalize from specific words in the lexicon
to all words in the language.
We argue that the present results are inconsistent with Costa and Santestebans (2004)
account. Instead, evidence suggested inhibition was observed in more restrictive contexts when
L2 learners planned speech relative to proficient bilinguals. Were we solely to consider the
magnitude of the inhibition, we might conclude that L2 learners rely heavily on inhibition.

However, one goal of this study was to move beyond investigating the presence or absence of
inhibition during speech planning. Instead, we attempted to characterize the precise properties of
regulatory mechanisms and their variation among different types of bilinguals. The present
methods, focusing on multiple dimensions of control, provided an elaborated account of
differences in language regulation that exist for less and more proficient bilinguals.
Interpreting mechanisms of language control through dynamics of interaction
The present results suggest that learners differ from proficient bilinguals because they
more focally engage inhibition for lexically-specific words. In addition, they suggest that
learners development of regulation is heavily influenced by whether L2 speech can be initiated
independently of the L1 translation equivalent. If we reconsider the pattern of lexically-specific
inhibition, it is important to note that inhibition was only observed when the repeated picture was
spoken first in the L2 and later in the L1. We suggest that this finding fits with the hypothesis
that L1 translations competed with L2 words during production. Like more recent studies
performed with proficient bilinguals, our findings suggest that the L1 translation route is still
operational despite the learners being beyond very beginning stages (e.g., Guo, Misra, Tam, &
Kroll, 2012, Thierry & Wu, 2007). However, we additionally suggest that L2 learners are
particularly more vulnerable to interference from L1 translations than proficient bilinguals
during L2 production (e.g., Costa, Caramazza, & Sebastian-Galls, 2000). The Revised
Hierarchical model (Kroll & Stewart, 1994) claims the L1 translation equivalent serves a critical
function of mediating access to the meaning of L2 words during early stages of learning. If
learners exploit the L1 to boost lexical access in the L2, then this creates an interesting dynamic
where learners are fundamentally open to the influence of the L1 but also dampening the
activation of the L1.

Though we sought ERP evidence to demonstrate how this is achieved, we observed ERP
modulations that were primarily consistent with priming. Additional research is needed to make
fine-grained predictions about the time course of competition between translation equivalents
during speech planning and the impact of graded differences in proficiency. We may assume,
given the very long-lasting time course of the lexically-specific effect that inhibition reflects
changes in the weighting in the connections from semantics to phonology rather than changes in
activation (e.g., Monsell, Matthews, & Miller, 1992; Oppenheim, Dell, & Schwartz, 2010).
Given that learners regained priming of lexically-specific words, to fully account for lexicallyspecific consequences, it will also be important to model the processes that enable L1 recovery.
It is interesting that when learners produced the L1 after the L2 a dissociation between
behavioral and neural measures emerged. Here, we reason that this dissociation is yet another
piece of evidence that different dynamics of cross-language activation arise for learners and are
critical for understanding how control manifests. This dissociation can be accommodated in a
model of speech production where the time course of cross-language activation not fixed (e.g.,
Hanulov, Davidson, & Indefrey, 2011; Kroll, Bobb, & Wodniecka, 2006). One might
additionally assume that since learners are so biased to produce words in the L1, they might not
be able to resolve competition from the L1 prior to response initiation. If not, then competition
will spill over onto articulation. In line with this idea, evidence for a prolonged time course of
cross language activation has been found for learners. Jacobs, Fricke, and Kroll (2015) observed
differences in the time course of cross-language interaction within two different groups of
intermediate L2 learners. One group of learners was in an immersion program where only the L2
could be spoken, and the other group was not immersed at the time of testing. Both groups of
learners demonstrated cognate effects in response latencies, but only the non-immersed

classroom learners demonstrated cognate effects that extended into articulation. This
counterintuitive finding was attributed to the classroom L2 learners differences limited
experience inhibiting the L1, relative to other groups that included even immersed learners. In
sum, the weak or absent effects in ERPs could reflect the tendency for language interaction (and
inhibitory control processes) to spill over onto articulation.
The influence of task factors on information processing during production
We investigated the scope and time course of language regulation when learners spoke
the L1 only. We predicted that learners who are not highly skilled at speaking the L2 would
function as if they were monolinguals, serving as a control for learners who actively spoke the
L2. In line with predictions, learners who were required to name pictures solely in the native
language demonstrated robust priming at the behavioral level. In particular, lexical priming was
revealed in the event-related potential record during the early N200. However, behavioral data
and ERPs also revealed unexpected costs. Since the context unambiguously required L1
production, these effects are probably due to factors associated with the task itself. This version
of the blocked picture naming task included a mixture of new pictures, new pictures from
repeated semantic categories, and literal repetitions to gauge effects at different levels of scope.
It is possible that more effortful or strategic processing emerged in response to task
manipulations of lexical and semantic category repetition. In a separate line of research, there has
been a long history of studying semantic interference effects that occur when specific category
exemplars are repeated in blocked and interleaved fashion (e.g., Belke, Meyer, & Damian, 2005;
Damian, Vogliocco, & Levelt, 2001; Schnur, Schwartz, Brecher, & Hodgson, 2006). There is
some evidence that in these tasks speakers recruit cognitive resources to resolve interference
(e.g., Crowther & Martin, 2014). Our paradigm did not implement exactly the same kind of

manipulations as the semantic blocking paradigm or cumulative semantic interference paradigm,

but it mirrored some of their properties. Because the current study did not set out to examine
these specific indices of interference, we cannot make strong claims about the origins of
interference observed in our task. Whatever the exact nature of the effect, the present results
suggest that task factors can induce changes in behavioral and neural processes during blocked
picture naming. In future studies, it will be important to identify contributions of the semantic
category and lexical repetition manipulations to blocked picture naming performance.
Future Directions
A recent line of research has demonstrated that consequences of bilingualism can be
observed when proficient bilinguals produce the L1 alone (e.g., Parker-Jones et al., 2011),
suggesting that there are very long-lasting inhibitory consequences. An unaddressed question
concerns whether long-lasting changes to the L1 induced by speaking the L2 are related to the
very-long lasting phenomenon of L1 suppression. Over time, L1 speech planning processes
might be restructured in bilinguals by virtue of frequently inhibiting the L1 to speak the L2.
Based on the Jacobs et al.s (2015) results, one prediction is that L2 learners who are immersed
in an L2 context (i.e., studying abroad) demonstrate inhibitory consequences for the L1
production even when it is spoken alone. This would provide additional evidence for the idea
that immersion experience accelerates the development of inhibitory control.
Conclusions
The blocked picture naming paradigm provided a powerful tool for investigating
processes of L2 speech planning in L2 learners. Despite learners clear challenge in speaking
single words in the L2, the requirement to speak the L2 produced robust changes to L1 speech
planning processes. In addition, fine-grained information about the scope of language regulation,

time course of regulation and L1 recovery was revealed. In summary, the blocked production
task provides a rich context for examining mechanisms of language regulation in learners.

References
Alario, F. X., Ferrand, L., Laganaro, M., New, B., Frauenfelder, U. H., & Segui, J. (2004).
Predictors of picture naming speed. Behavior Research Methods, Instruments, &
Computers, 36, 140-155.
Balota, D. A., Cortese, M. J., Hutchison, K. A., Neely, J. H., Nelson, D., Simpson, G. B., &
Treiman, R. (2002). The English Lexicon Project: A web-based repository of descriptive
and behavioral measures for 40,481 English words and nonwords. Washington
University. Online: http://elexicon. wustl. edu.
Barry, C., Hirsh, K. W., Johnston, R. A., & Williams, C. L. (2001). Age of acquisitions, word
frequency, and the locus of repetition priming of picture naming. Journal of Memory and
Language, 44, 350-375. doi:10.1006/jmla.2000.2743
Belke, E., Meyer, A. S., & Damian, M. F. (2005). Refractory effects in picture naming as
assessed in a semantic blocking paradigm. The Quarterly Journal of Experimental
Psychology, 58, 667-692.
Bobb, S. C., & Wodniecka, Z. (2013). Language switching in picture naming: What asymmetric
switch costs (do not) tell us about inhibition in bilingual speech planning. Journal of
Cognitive Psychology, 25, 568585. doi:10.1080/20445911.2013.792822
Branzi, F. M., Martin, C. D., Abutalebi, J., & Costa, A. (2014). The after-effects of bilingual
language production. Neuropsychologia, 52, 102116.
doi:10.1016/j.neuropsychologia.2013.09.022

Cave, C. B. (1997). Very long-lasting priming in picture naming. Psychological Science, 8, 322325. doi: 10.1111/j.1467-9280.1997.tb00446.x
Cave, C. B. & Squire, L. R. (1992). Intact and long-lasting repetition priming in amnesia.
Journal of Experimental Psychology: Learning, Memory and Cognition, 18, 509-520.
doi:10.1037/0278-7393.18.3.509
Christoffels, I. K., Firk, C., & Schiller, N. O. (2007). Bilingual language control: An eventrelated brain potential study. Brain research, 1147, 192-208.
doi:10.1016/j.brainres.2007.01.137
Coltheart, M. (1981). The MRC psycholinguistic database. The Quarterly Journal of
Experimental Psychology: Human Experimental Psychology, 33, 497-505.
doi:10.1080/14640748108400805
Costa, A., Caramazza, A., & Sebastian-Galles, N. (2000). The cognate facilitation effect:
implications for models of lexical access. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 26, 12831296. doi:10.1037/0278-7393.26.5.1283
Costa, A., & Santesteban, M. (2004). Lexical access in bilingual speech production: Evidence
from language switching in highly proficient bilinguals and L2 learners. Journal of
Memory and Language, 50, 491511. doi:10.1016/j.jml.2004.02.002
Costa, A., Santesteban, M., & Ivanova, I. (2006). How do highly proficient bilinguals control
their lexicalization process? Inhibitory and language-specific selection mechanisms are
both functional. Journal of Experimental Psychology: Learning, Memory, and Cognition,
32, 10571074. doi:10.1037/0278-7393.32.5.1057

Crowther, J. E., & Martin, R. C. (2014). Lexical selection in the semantically blocked cyclic
naming task: the role of cognitive control and learning. Frontiers in human
neuroscience,8.
Cuetos, F., Ellis, A. W., Alvarez, B. (1999). Naming times for the Snodgrass and Vanderwart
pictures in Spanish. Behavior Research Methods, Instruments, & Computers, 31, 650658.
Damian, M. F., Vigliocco, G., & Levelt, W. J. (2001). Effects of semantic context in the naming
of pictures and words. Cognition, 81, B77-B86.
Ecker, U. K., & Zimmer, H. D. (2009). ERP evidence for flexible adjustment of retrieval
orientation and its influence on familiarity. Journal of Cognitive Neuroscience, 21,
1907-1919. doi:10.1162/jocn.2009.21135.
Falkenstein, M., Hoormann, J., & Hohnsbein, J. (1999). ERP components in Go/Nogo tasks and
their relation to inhibition. Acta Psychologica, 101, 267-291. doi:10.1016/S00016918(99)00008-6
Franklin, M. S., Dien, J., Neely, J. H., Huber, E., & Waterson, L. D. (2007). Semantic priming
modulates the N400, N300, and N400RP. Clinical Neurophysiology, 118, 1053-1068.
Green, D. (1998). Mental control of the bilingual lexico-semantic system, Bilingualism:
Language and Cognition, 1, 6781. doi:10.1017/S1366728998000133
Greenhouse, S. W., & Geisser, S. (1959). On methods in the analysis of profile data.
Psychometrika, 24, 95-112. doi:10.1007/BF02289823
Hanulov, J., Davidson, D. J., & Indefrey, P. (2011). Where does the delay in L2 picture naming
come from? Psycholinguistic and neurocognitive evidence on second language word

production. Language and Cognitive Processes, 26, 902-934.

doi:10.1080/01690965.2010.509946.
Henson, R. N., Rylands, A., Ross, E., Vuilleumeir, P., & Rugg, M. D. (2004). The effect of
repetition lag on electrophysiological and haemodynamic correlates of visual object
priming. Neuroimage, 21, 1674-1689.
Hoshino, N., & Thierry, G. (2011). Language selection in bilingual word production:
Electrophysiological evidence for cross-language competition. Brain Research, 1371,
100109. doi:10.1016/j.brainres.2010.11.053
Jackson, G. M., Swainson, R., Cunnington, R., & Jackson, S. R. (2001). ERP correlates of
executive control during repeated language switching. Bilingualism: Language and
Cognition, 4, 169178. doi:10.1017/S1366728901000268
Jacobs, A., Fricke, M. and Kroll, J. F. (2015), Cross-Language Activation Begins During Speech
Planning and Extends Into Second Language Speech. Language Learning.
doi: 10.1111/lang.12148.
Kroll, J. F., & Bobb, Susan C, Misra, Maya, Guo, T. (2008). Language selection in bilingual
speech: Evidence for inhibitory processes. Acta Psychologica, 128, 416430.
doi:10.1016/j.actpsy.2008.02.001.
Kroll, J., Michael, E., Tokowicz, N., & Dufour, R. (2002). The development of lexical fluency in
a second language. Second Language Research, 18, 137171.
doi:10.1191/0267658302sr201oa
Kroll, J. F., & Stewart, E. (1994). Category interference in translation and picture naming:

Evidence for asymmetric connections between bilingual memory representations. Journal

of Memory and Language, 33, 149-174. doi:10.1006/jmla.1994.1008
Kucera, H., & Francis, W. N. (1967). Computational analysis of present-day American English.
Providence, RI: Brown University Press.
Kutas, M. & Hillyard, S.A. An electrophysiological probe of incidental semantic association.
(1989). Journal of Cognitive Neuroscience, 1, 3849.
McLaughlin, J., Osterhout, L., & Kim, A. (2004). Neural correlates of second-language word
learning: Minimal instruction produces rapid change. Nature neuroscience, 7, 703704. doi:10.1038/nn1264
McPherson, W. B., & Holcomb, P. J. (1999). An electrophysiological investigation of semantic
priming with pictures of real objects. Psychophysiology, 36, 53-65.
McRae, K., Cree, G. S., Seidenberg, M. S., & McNorgan, C. (2005). Semantic feature production
norms for a large set of living and nonliving things. Behavior Research Methods, 37,
547-559. doi:10.3758/BF03192726
Midgley, K. J., Holcomb, P. J., & Grainger, J. (2009). Masked repetition and translation priming
in second language learners: A window on the time-course of form and meaning
activation using ERPs. Psychophysiology, 46, 551565. doi:10.1111/j.14698986.2009.00784.x
Miozzo, M., Pulvermller, F., & Hauk, O. (2014). Early parallel activation of semantics and
phonology in picture naming: Evidence from a multiple linear regression MEG study.
Cerebral Cortex, bhu137.
Misra, M., Guo, T., Bobb, S. C., & Kroll, J. F. (2012). When bilinguals choose a single word to

speak: Electrophysiological evidence for inhibition of the native language. Journal of

Memory and Language, 67, 224237. doi:10.1016/j.jml.2012.05.001
Mitchell, D. B. & Brown. A. S. (1988). Persistent repetition priming in picture naming and its
dissociation from recognition memory. Journal of Experimental Psychology: Learning,
Memory, & Cognition, 14, 213-222. doi.org/10.1037/0278-7393.14.2.213
Monsell, S., Matthews, G. H., & Miller, D. C. (1992). Repetition of lexicalization across
languages: A further test of the locus of priming. The Quarterly Journal of Experimental
Psychology, 44, 763-783.
Nieuwenhuis, S., Yeung, N., van den Wildenberg, W. & Ridderinkhof, K. R. (2003).
Electrophysiological correlates of anterior cingulate function in a go/nogo task: Effects of
response conflict and trial type frequency. Cognitive, Affective, & Behavioral
Neuroscience, 3, 17-26. doi:10.3758/CABN.3.1.17
Oppenheim, G. M., Dell, G. S., & Schwartz, M. F. (2010). The dark side of incremental learning:
A model of cumulative semantic interference during lexical access in speech production.
Cognition, 114, 227-252.
Parker Jones, ., Green, D. W., Grogan, A., Pliatsikas, C., Filippopolitis, K., Ali, N., Price,
C. J. (2012). Where, when and why brain activation differs for bilinguals and
monolinguals during picture naming and reading aloud. Cerebral Cortex, 22, 892
902.doi: 902.10.1093/cercor/bhr161.
Philipp, A. M., Gade, M., & Koch, I. (2007). Inhibitory processes in language switching:
Evidence from switching language-defined response sets. European Journal of Cognitive

Psychology, 19, 395416. doi:10.1080/09541440600758812

Philipp, A. M., & Koch, I. (2009). Inhibition in language switching: what is inhibited when
switching between languages in naming tasks? Journal of Experimental Psychology:
Learning, Memory, and Cognition, 35, 11871195. doi:10.1037/a0016376
Schnur, T. T., Schwartz, M. F., Brecher, A., & Hodgson, C. (2006). Semantic interference during
blocked-cyclic naming: Evidence from aphasia. Journal of Memory and Language, 54,
199-227.
Snodgrass, J.G. & Vanderwart, M. (1980). A standardized set of 260 pictures: norms for name
agreement, image agreement, familiarity, and visual complexity. Journal of Experimental
Psychology: Human Learning and Memory, 6, 174-215. doi:10.1037/0278-7393.6.2.174
Snodgrass, J. G., & Yuditsky, T. (1996). Naming times for the Snodgrass and Vanderwart
pictures. Behavior Research Methods, Instruments & Computers, 28, 516-536.
doi:10.3758/BF03200741
Strijkers, K., Costa, A., & Thierry, G. (2009). Tracking lexical access in speech production:
electrophysiological correlates of word frequency and cognate effects. Cerebral cortex,
bhp153. doi: 10.1093/cercor/bhp153
Strijkers, K., Holcomb, P. J., & Costa, A. (2011). Conscious intention to speak proactively
facilitates lexical access during overt object naming. Journal of Memory and Language,
65, 345-362. doi:10.1016/j.jml.2011.06.002
Sunderman, G., & Kroll, J. F. (2006). First language activation during second language lexical
processing. Studies in Second Language Acquisition, 28, 387422.

doi:10.1017/S0272263106060177
Thierry, G., & Wu, Y. J. (2007). Brain potentials reveal unconscious translation during foreignlanguage comprehension. Proceedings of the National Academy of Sciences, 104, 1253012535.
Tokowicz, N., Michael, E. B., & Kroll, J. F. (2004). The roles of study-abroad experience and
working-memory capacity in the types of errors made during translation. Bilingualism:
Language and Cognition, 7, 255272. doi:10.1017/S1366728904001634
Van Assche, E., Duyck, W., & Gollan, T. (2013). Whole-language and item-specific control in
bilingual language production. Journal of Experimental Psychology: Memory, Language
and Cognition, 39, 1781-1792. doi:10.1037/a0032859
Van Casteren, M., & Davis, M. H. (2007). Match: A program to assist in matching the conditions
of factorial experiments. Behavior Research Methods, 39, 973-978.
Verhoef, K., Roelofs, A., & Chwilla, D. J. (2009). Role of inhibition in language switching:
Evidence from event-related brain potentials in overt picture naming. Cognition, 110, 84
99. doi:10.1016/j.cognition.2008.10.013

Appendix A: Language history and the self-assessed proficiency for all participants
Table 1
Participant Group
Variable
Age

Experiment 1

Experiment 2

21.4 (1.9)

21.2 (1.5)

5.1

2.4

Semesters of
Spanish
Operation Span
Flanker Effect
L1 Self-rated
Proficiency
Reading

88 (9)
43 (23)

90 (8)
43 (27)

9.89 (0.32)

9.89 (0.32)

Writing

9.67 (0.59)

9.83 (0.38)

Speaking

9.83 (0.38)

9.94 (0.24)

Spelling

9.39 (0.50)

9.67 (0.48)

Listening

9.94 (0.23)

9.94 (0.24)

L2 Self-rated
Proficiency
Reading

7.56 (1.34)

6.94 (1.35)

Writing

7.06 (1.51)

6.39 (1.42)

Speaking

6.83 (1.62)

5.94 (1.83)

Spelling

7.94 (1.71)

7.39 (1.85)

Listening

7.61 (1.50)

7.01 (1.70)

Note:
a
n=18.
b
n=19.
c. Reported in years
d. Number of correctly recalled words out of 60 possible
e. Self-ratings of proficiency (on a 1-10 Likert scale) in both languages. Standard deviations are indicated
between parentheses.
f. Self-ratings of proficiency are based on a scale where 10 indicates the highest level of fluency.
g. Flanker effect was calculated in milliseconds.
Standard deviations are indicated within parenthesis.

Appendix B. Properties of the materials.

Table 2
Characteristics of Pictures and Picture Names
Variable
Blocks 1 and 2
New Pictures:

New Pictures:

New Pictures:

p-value
(F-test)

(and respective

From Block 1

From Block 2

From Entirely

Lexical Repetitions)

Semantic Categories

Semantic Categories

New Categories

Syllablesc

1.4 (.55)

1.5 (.74)

1.5 (.94)

1.7 (.70)

ns

Frequencya

21.2 (33.5)

22.4 (35.2)

28.0 (28.6)

21.3 (26.4)

ns

Familiarityb

6.9 (2.0)

6.6 (1.4)

7.3 (1.3)

6.7 (1.5)

ns

Featuresd

14.8 (4.2)

15 (3.3)

14.7 (2.8)

14.7 (3.0)

ns

Notes. Blocks 1 and 2 were the same items but counterbalanced across participants, therefore means are
a

averaged together. Kuera & Francis (1967); Per million words. Data from the MRC Psycholinguistic
c

database (Coltheart, 1981). Data from the English Lexicon Project database (Balota et al., 2002). McRae
et al. (2005); Number of taxonomic semantic featuresd. New pictures from Block1 and Block 2 Semantic
Categories were included in the materials for both Experiments 1 and 2, but the analyses of these data are
not reported here in the present paper.

Table 3.
Model Summary: Experiment 1. Mean response latency and accuracy per Block and Type.
Latency

Accuracy

F1 Analysis

F(1,17)=55.56, p<.001

F(1,17)=106.01, p<.001

F2 Analysis

F(1,53)=73.33 p<.001

F(1,53) =96.78, p <.001

F1 Analysis

F(3,51)=3.99, p<.01

F<1

F2 Analysis

F<1

F<1

F1 Analysis

F(2,34)=3.43, p<.05

F(2,34)=2.64, p=.09

F2 Analysis

F<1

F<1

F1 Analysis

F(2,34)=9.79, p<.001

F(2,34)=3.81, p=.06

F2 Analysis

F(2,106)=15.59, p<.001

F(2,106)=8.17, p<.05

F1 Analysis

F(4,68)=2.49, p=.05

F<1

F2 Analysis

F<1

F<1

Block 1 vs. Block 2

Block

Block 1 vs. New Pictures

Block

New vs. Lexically-identical Repetitions

Block

Type

Block x Type

Table 4.
Model Summary: Experiment 1. ERP Analysis Comparing Block 1 vs. Block 2.
N200 component (200-270 ms)

N400 component (350-450 ms)

Midline
Block

F-value, p-value
F <1

F-value, p-value
F <1

Electrode

F(3, 51)=31.02, p=.001

Block * Electrode

F(3,51 ) =25.07, p <.001

F <1

Medial-Lateral
Block

F <1

F <1

Electrode

F(4, 68)=18.19, p <.001

F(4, 68)=34.84, p <.001

Hemisphere

F(1,17)=5.81, p <.05

F <1

Block * Electrode

F <1

F <1

Block * Hemisphere

F <1

F <1

Block * Electrode * Hemisphere

F <1

F <1

Lateral-Lateral
Block

F <1

F <1

Electrode

F(2, 34)=17.70, p <.001

F(2, 34)=1478.31, p<.001

Hemisphere

F(1,17 )=14.77, p <.001

F(1,17 )=229.56, p <.001

Block * Electrode

F <1

F(2, 34)=6.87, p <.05

Block * Hemisphere

F <1

F <1

Block * Electrode * Hemisphere

F <1

F <1

F(3, 51)=2.83, p=.09

Table 5.
Model Summary: Experiment 1. ERP Analysis comparing new vs. lexical repetitions.
N200 component (200-270 ms)

N400 component (350-450 ms)

Midline
Block

F-value, p-value
F <1

F-value, p-value

Electrode

F(3, 51)= 22.09, p<.001

F(3,51 ) =28.36, p <.001

Type

F(2, 34)=15.40, p<.001

Block * Electrode

F <1

F(2,34 ) =10.16, p =.001

F <1

Block * Type

F <1

F <1

Type * Electrode

F(6, 102)= 3.36, p<.05

F(6,102 ) =3.13, p =.01

Medial-Lateral
Block

F <1

F <1

Electrode

F(3, 51)= 26.76, p<.001

F(4,68 ) =24.50, p <.001

Hemisphere

F <1

F<1

Type

F(2, 34)=8.92, p=.001

F(2,34 ) =7.64, p <.01

Block * Electrode

F <1

F <1

Block * Hemisphere

F <1

F <1

Block * Type

F <1

F <1

Block * Electrode * Hemisphere

F <1

F <1

Block * Electrode * Type

F <1

F <1

Lateral-Lateral
Block

F <1

F <1

Electrode

F(2, 34)=25.08, p <.001

F(2, 34)=25.27, p <.001

Hemisphere

F(1,17 )=10.07, p <.01

F <1

Type

F(2, 34)=9.49, p =.001

F(2, 34)=2.55, p =.09

Block * Electrode

F <1

F <1

Block * Hemisphere

F <1

F(2, 34)=3.36, p =.05

Block * Type

F <1

F <1

Block * Electrode * Hemisphere

F <1

F <1

Block * Electrode * Type

F <1

F <1

F(2,34 ) =2.62, p =.09

Table 6.
Model Summary: Experiment 1. ERP Analysis comparing L1 naming in Block 1 to new pictures.
N200 component (200-270 ms)

N400 component (350-450 ms)

Midline
Block

F-value, p-value
F <1

F-value, p-value
F <1

Electrode
Block * Electrode

F(3, 51)= 21.55, p=.001

F <1

F(3,51 ) =28.19, p <.001

F <1

Medial-Lateral
Block

F <1

F <1

Electrode

F(4, 68)=23.89, p <.001

F(4, 68)=22.76, p <.001

Hemisphere

F <1

F<1

Block * Electrode

F <1

F <1

Block * Hemisphere

F <1

F <1

Block * Electrode * Hemisphere

F <1

F <1

Lateral-Lateral
Block

F(3, 51)=2.64, p =.06

F(3, 51)=10.20, p <.001

Electrode

F(2, 34)=23.40, p <.001

F(2, 34)=26.45, p <.001

Hemisphere

F(1,17 )=9.57, p <.01

F <1

Block * Electrode

F(2, 34)=6.87, p <.05

F <1

Block * Hemisphere

F <1

F <1

Block * Electrode * Hemisphere

F <1

F(6, 102)=3.75, p =.01

Table 7.
Model Summary: Experiment 2. Mean response latency and accuracy per Block and Type.
Latency

Accuracy

F1 Analysis

F <1

F <1

F2 Analysis

F(1,53)=12.75, p<.01

F(1,53)=12.75, p =.001

F1 Analysis

F(3,54)=2.66, p=.06

F(3,54)=3.38, p<.05

F2 Analysis

F <1

F <1

F1 Analysis

F <1

F <1

F2 Analysis

F <1

F <1

F1 Analysis

F(2,36)=32.93, p<.001

F(2,36)=28.21, p<.001

F2 Analysis

F(2,106)=32.04, p<.001

F(2,106)=16.90, p<.001

F1 Analysis

F <1

F(4,72)=2.12, p=.09

F2 Analysis

F <1

F <1

Block 1 vs. Block 2

Block

Block 1 vs. New Pictures

Block

New vs. Lexically-identical Repetitions

Block

Type

Block x Type

Table 8.
Model Summary: Experiment 2. ERP Analysis Comparing Block 1 vs. Block 2.
N200 component (200-270
ms)

N400 component (350-450 ms)

Midline
Block

F-value, p-value

F-value, p-value

F <1

F <1

Electrode
Block * Electrode

F(3, 54)=38.06, p<.001

F <1

F(3,54 ) =27.84, p <.001

F <1

Medial-Lateral
Block

F <1

F <1

Electrode

F(4, 72)=35.57, p <.001

F(4, 72)=28.51, p <.001

Hemisphere

F <1

F(1,18)=3.10, p =.10

Block * Electrode

F <1

F(4,72)=6.04, p <.05

Block * Hemisphere

F(1, 18)=5.43, p <.05

F <1

Block * Electrode *
Hemisphere

F <1

F <1

Lateral-Lateral
Block

F <1

F <1

Electrode

F(2, 36)=49.81, p<.001

F(2, 34)=45.80, p <.001

Hemisphere

F(1,18 )=12.36, p =.01

F(1,17 )=16.76, p =.001

Block * Electrode

F(2, 36)=3.88, p =.06

F <1

Block * Hemisphere

F(1,18 )=8.29, p =.01

F(1,18 )=3.86, p =.07

Block * Electrode *
Hemisphere

F(2,36 )=4.56, p <.05

F <1

Table 9.
Model Summary: Experiment 2. ERP Analysis comparing new vs. lexical repetitions.
N200 component (200-270 ms)

N400 component (350-450 ms)

Midline
Block

F-value, p-value
F <1

F-value, p-value
F <1

Electrode

F(3, 54)= 21.91, p<.001

F(3, 54) =37.68, p <.001

Type

F(2, 36)=7.10, p<.01

Block * Electrode

F <1

F(2, 36) =3.21, p =.05

F <1

Block * Type

F <1

F <1

Type * Electrode

F <1

F <1

Medial-Lateral
Block

F <1

F <1

Electrode

F(3, 54)= 31.23, p<.001

F(4,72 ) =45.24, p <.001

Hemisphere

F <1

F(1,18 ) =4.66, p <.05

Type

F(2, 36)=5.53, p<.01

F <1

Block * Electrode

F <1

F <1

Block * Hemisphere

F <1

F <1

Block * Type

F <1

F <1

Block * Electrode * Hemisphere

F <1

F <1

Block * Electrode * Type

F <1

F <1

Lateral-Lateral
Block

F <1

F <1

Electrode

F(2, 36)=46.45, p <.001

F(2, 36)=53.66, p <.001

Hemisphere

F(1,18 )=11.60, p <.01

F(1,18 )=23.45, p <.001

Type

F(2, 36)=6.27, p <.01

F <1

Block * Electrode

F <1

F <1

Block * Hemisphere

F <1

F <1

Block * Type

F <1

F <1

Block * Electrode * Hemisphere

F <1

F(4, 72)=2.55, p =.07

Block * Electrode * Type

F <1

F <1

Table 10.
Model Summary: Experiment 2. ERP Analysis comparing L1 naming in Block 1 to new pictures.
N200 component (200-270 ms)

N400 component (350-450 ms)

Midline
Block

F-value, p-value
F(3, 54)= 6.57, p<.05

F-value, p-value
F <1

Electrode
Block * Electrode

F(3, 54)= 34.09, p<.001

F <1

F(3,54 ) =41.21, p <.001

F <1

Medial-Lateral
Block

F(3, 54)= 2.39, p=.08

F <1

Electrode

F(4, 72)=33.56, p <.001

F(4, 72)=49.92, p <.001

Hemisphere

F <1

F(3, 54)= 4.21, p=.06

Block * Electrode

F(3, 54)= 2.09, p=.09

F <1

Block * Hemisphere

F <1

F <1

Block * Electrode *
Hemisphere

F <1

F <1

Lateral-Lateral
Block

F <1

F <1

Electrode

F(2, 36)=51.39, p <.001

F(2, 34)=68.74, p <.001

Hemisphere

F(1,18 )=15.50, p =.001

F(1,18 )=24.86, p <.001

Block * Electrode

F <1

F <1

Block * Hemisphere

F(1,18 )=4.56, p <.01

F(1,18 )=2.98, p <.05

Block * Electrode *
Hemisphere

F <1

F(6, 102)=2.46, p =.06

Appendix C: ERP figures of critical electrodes within the midline, left, and right lateral columns.
Experiment 1. Grand average ERPs for the comparison of New and Repeated Pictures (lexical repetitions from Blocks 1 and 2).
Negative is plotted up.

-10

F3

FZ

-10

F4
-10

-5

-5

-5

10
-150 -50
-10

50

150 250 350 450 550

10
-150 -50

50

C3

150 250 350 450 550

CZ

10
-150 -50

50

-10

150 250 350 450 550

C4

-10
-5

-5

-5

0
5
10
-150 -50

-10

5
50 150 250 350 450 550

P3

10
-150 -50

50

150 250 350 450 550

PZ

-10

10
-150 -50

150 250 350 450 550

P4

-10
-5

-5

-5

50

10
-150 -50

50 150 250 350 450 550

10
-150 -50

50

150 250 350 450 550

10
-150 -50

50

150 250 350 450 550

Experiment 1. Grand average ERPs for the comparison of Block 1 and New Pictures from Blocks 3-8. Negative is plotted up.

F4

FZ

F3
-10

-10

-10

-5

-5

-5

, 0

10
-150 -50

-10

50 150 250 350 450 550

C3

-5
0
5
10
-150 -50

10
-150 -50 50 150 250 350 450 550

50 150 250 350 450 550

10
-150 -50 50 150 250 350 450 550

CZ

C4

-10

-10

-5

-5

10
-150 -50

P3

50 150 250 350 450 550

10
-150 -50 50 150 250 350 450 550

P4

PZ

-10

-10

-10

-5

-5

-5

10
-150 -50

50 150 250 350 450 550

10
-150 -50 50 150 250 350 450 550

10
-150 -50 50 150 250 350 450 550

Experiment 2. Grand average ERPs for the comparison of New and Repeated Pictures (lexical repetitions from Blocks 1 and 2).
Negative is plotted up.
F3

-10

FZ

-10

-5

-5

F4

-10
-5
0
5

5
10
-150 -50

50

150 250 350 450 550

C3

-10

10

10
-150 -50

50

150 250 350 450 550

15
-150 -50

50

150 250 350 450 550

-10

CZ

C4

-10

-5
-5

-5
0

0
5

5
10
-150 -50

5
10
-150 -50
50

P3

-10

50

150 250 350 450 550

150 250 350 450 550

PZ

-10

10
-150 -50

-5

-5

50

150 250 350 450 550

10
-150 -50

50

150 250 350 450 550

150

10
-150 -50

250

350

450

550

P4

-10

-5

10
-150 -50

50

50

150 250 350 450 550

Experiment 2. Grand average ERPs for the comparison of Block 1 and New Pictures from Blocks 3-8. Negative is plotted up.

F3

F4

FZ

-10

-10

-10

-5

-5

-5

10
-150 -50 50 150 250 350 450 550

10
-150 -50 50 150 250 350 450 550

C3

10
-150 -50 50 150 250 350 450 550

C4

CZ

-10

-10
-10

-5

-5

-5

10
-150 -50

50 150 250 350 450 550

10
-150 -50 50 150 250 350 450 550

P3
-10

10
-150 -50

-10

-5

-5

-5

10
-150 -50 50 150 250 350 450 550

10
-150 -50 50 150 250 350 450 550

150 250 350 450 550

P4

PZ
-10

50

10
-150 -50

50 150 250 350 450 550

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure Captions
Figure 1. Mean response latency (in milliseconds) of L2 learners naming pictures during Blocks 1 and 2 in Experiments 1 and 2. Note
that in Experiment 1, Block 2 was performed in the L2 (Spanish). Error bars show one standard error.
Figure 2. Magnitude of the priming (in milliseconds): L2 learners naming lexical repetitions in Experiment 1 and 2. Left panel:
Learners who named in Spanish in Experiment 1. Right panel: Learners who named in L1 only in Experiment 2. Error bars show one
standard error.
Figure 3. Mean Response Latency (in milliseconds): Mean response latency (in milliseconds) of L2 learners naming new pictures
during Block 1 and the later blocks (3-8) in Experiment 1. Error bars are one standard error.
Figure 4. (A) Grand average ERP waveform for electrode site Fz for new and lexical repetition trials from Block 1 and Block in
Experiment 1. (B) Grand average ERP waveform for electrode site Fz for new and lexical repetition trials from Block 1 and Block 2 in
Experiment 2.
Figure 5. (A) Grand average ERP waveform for electrode site Fz for Block 1 (solid black line), New pictures in Blocks 3 and 4 (long
dashed line), New pictures in Blocks 5 and 6 (small dotted line), and New pictures in Blocks 7 and 8 (short dashed line) in Experiment
1. (B) The same conditions are plotted for learners who participated in Experiment 2. The time windows 200270 ms and 350450
ms were used for statistical analyses. Negative is plotted up.