Novice Learners, Longitudinal Designs, and

Event-Related Potentials: A Means for Exploring the

Neurocognition of Second Language Processing
Lee Osterhout, Judy Mclaughlin, Ilona Pitkänen, Cheryl Frenck-Mestre,
Nicola Molinaro

To cite this version:

Lee Osterhout, Judy Mclaughlin, Ilona Pitkänen, Cheryl Frenck-Mestre, Nicola Molinaro. Novice
Learners, Longitudinal Designs, and Event-Related Potentials: A Means for Exploring the Neurocogni-
tion of Second Language Processing. Language Learning, Wiley, 2006, 56, pp.199-230. �hal-00380711�

HAL Id: hal-00380711

https://hal.archives-ouvertes.fr/hal-00380711
Submitted on 4 May 2009

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 Osterhout et al.
1

Novice learners, longitudinal designs, and event-related potentials:

A means for exploring the neurocognition of second-language processing

Lee Osterhout, Judith McLaughlin, Ilona Pitkänen

University of Washington

Cheryl Frenck-Mestre
Centre National de la Recherche Scientifique
Université de Provence

Nicola Molinaro
Università delgli Studi di Padova

Please address correspondence to:

Lee Osterhout
Department of Psychology
Box 351525
University of Washington
Seattle, WA 98195
USA

email: [email protected]
phone: (206) 543-4177
 Osterhout et al.
2
Abstract

Research on the neurobiology of second-language (L2) learning has historically focused

on localization questions and relied on cross-sectional designs. Here, we describe an

alternative paradigm involving longitudinal studies of adult, novice learners who are

progressing through an introductory sequence of classroom-based L2 instruction. The

goal of this paradigm is to determine how much L2 exposure is needed before learners

incorporate L2 knowledge into their on-line comprehension processes, as reflected in

scalp-recorded event-related brain potentials (ERPs). Our preliminary studies show that

some, but not all, aspects of the L2 (including lexical and morphosyntactic aspects) are

incorporated into the comprehension system after remarkably little L2 instruction. We

discuss the benefits of this paradigm while acknowledging the limitations and potential

difficulties associated with it.

 Osterhout et al.
3

One could argue that two historical facts have impeded progress in understanding

how a second language (L2) is instantiated in a language learner’s brain. The first

historical fact is the emphasis on the “where” question, that is, the question of which

parts of the brain are involved in using the L2. Although the “where” question is a valid

and interesting one, the focus on this question has led to neglect of other questions that

might prove to be equally important. The second historical fact concerns how researchers

have dealt with two sources of variability. The first type, variability among L2 learners,

introduces potential confounding variables that are well understood but rarely adequately

controlled for. The second type, variability within a single learner over time, is a

potentially crucial source of evidence that has been largely ignored. Our goal in this

commentary is to argue for an alternative paradigm for studying the neurocognition of L2

learning and usage, in which the focus is on how the processes underlying comprehension

of the L2 change with increasing L2 exposure or proficiency. The paradigm minimizes

the troublesome sources of between-learner variability while maximizing sensitivity to

theoretically crucial within-learner variability over time.

Two primary methods have been used to identify the neural structures involved in

using an L2. One method involves studying aphasic bilinguals. If a patient’s facility with

two languages can be damaged independently, then one reasonable interpretation is that

the two languages are represented independently in the speaker’s brain. A second

method involves the use of neuroimaging techniques (e.g., functional magnetic resonance

imaging, or fMRI) to contrast L1 and L2 localization within the brain. These tools

provide a measure of the brain’s metabolic activity in neurologically intact individuals,

 Osterhout et al.
4
and allow researchers to determine which areas of the brain are metabolically most

active, given some task. Evidence that different areas of the brain become activated

when a bilingual uses her first or second language could, again, be interpreted to mean

that different neural areas are involved in using the two languages.

However, nearly every pattern of language loss and recovery (loss of a first

language with preservation of a second, or equal loss of both languages, etc.) has been

reported in bilingual aphasics. Similarly, whereas a few neuroimaging studies seem to

indicate a (partial) neuroanatomical separation of the two languages (Kim et al., 1997;

Perani et al., 1998), others seem to indicate a complete overlap in the brain areas

activated by the two languages (Chee, Tan, & Thiel, 1999; Klein, Milner, Zatorre, Zhao,

& Nikelski, 1999). It is reasonable to assume that the lack of consistency across

experiments reflects, in part, uncontrolled subject variability (variability in learner

motivation, the type and nature of L2 exposure, the age of L2 acquisition, etc), a problem

that has vexed much of the L2 literature (Grosjean, 1998). Furthermore, the range of

inferences that these methods permit is quite limited. For example, language dysfunction

might result from damage to white matter tracts that course through the area, rather than

from damage to the gray matter at the lesion site. A complete overlap in fMRI activation

patterns for an L1 and an L2 does not necessarily imply that the two languages are

processed in exactly the same way; and a partial overlap in activation patterns gives us

little idea of how processing of the two languages might differ.

Rather than focus on the “where” question, the paradigm we have in mind focuses

on the processes that recognize words and derive the structure and meaning of sentences.

More specifically, the paradigm is designed to help us learn more about how these
 Osterhout et al.
5
processes (and the neural systems that underlie them) change over time, with increasing

L2 exposure or proficiency. This paradigm minimizes between-subjects variability by

longitudinally studying novice L2 learners who are progressing through their first years

of classroom L2 instruction. In this design, the learners are highly similar in a priori L2

proficiency (having none) and experience a highly similar learning environment.

Because each learner acts as his or her own control, many of the subject variables that

can potentially confound L2 studies have been eliminated within this design. By also

including native speakers of the L2 in the study, we are able to compare the beginning

and intermediate states of learning to the end state of native proficiency.

The longitudinal design also maximizes sensitivity to the within-subject variation

that is the primary focus of our research program. Our goal is to identify changes in brain

activity (and in particular activity that reflects on-line L2 processing) that accompany the

earliest stages of L2 learning. We do so by recording event-related brain potentials

(ERPs) from the scalp while learners read L2 words or sentences. ERPs provide an on-

line, millisecond-by-millisecond record of the brain’s electrical activity during language

comprehension, and therefore (unlike fMRI and PET) have the necessary temporal

resolution to isolate specific language comprehension processes in time. Furthermore,

ERPs are multidimensional, varying in polarity, timing, morphology, and scalp

distribution; this multidimensionality might in theory provide ERPs with a differential

sensitivity to at least some of the processing steps underlying comprehension.

In fact, ERPs do seem to be differentially sensitive to events occurring at distinct

levels of linguistic analysis (for a review, see Osterhout, McLaughlin, Kim, & Inoue,

2004). One ERP response, the N400 component, is sensitive to properties of words, both
 Osterhout et al.
6
when they appear in isolation and when they appear in a linguistic context. A second

component, the P600, is sensitive to the grammatical well-formedness of sentences.

Although the precise cognitive and neural events underlying these effects are not known,

their existence and particular sensitivities can be exploited to learn a great deal about

language comprehension and (we hope to demonstrate here) second language learning.

We will demonstrate that although ERPs may not be tremendously informative with

respect to the “where” question, they are in some instances extremely useful for

ascertaining what is happening, and when it happens. The method’s sensitivity to the

“what” question follows from the fact that the language-relevant ERP effects are quite

specific in their sensitivities, thereby permitting fairly specific inferences about what is

happening at that moment in processing. The method’s sensitivity to the “when” question

encompasses two time scales: the small intervals of time (tens and hundreds of

milliseconds) over which a word or sentence is processed, and the large intervals of time

(weeks, months, and years) over which a person progresses from no competence with a

language to increasing competence.

In the next section, we briefly review what has been learned about the relationship

between language processing and ERPs, discussing in particular problems that might

come up when applied to L2 research. In the subsequent sections, we provide examples

of how this knowledge, when combined with the research paradigm described above, can

be used to gain some new insights about what happens during L2 learning and

processing, and when it happens.

 Osterhout et al.
7
ERPs as tools for studying first and second language processing

Kutas and colleagues (Kutas & Hillyard, 1980) were the first to demonstrate that

semantically anomalous words (e.g., “He spread his warm bread with socks”) elicit an

increase in the amplitude of the N400 component, a negative-going wave that peaks at

about 400 ms (Fig 1a). This result is observed regardless of the position of the anomalous

word within the sentence or the modality of input (visual vs. auditory; Osterhout et al

1993). Subsequent research has shown that N400 amplitude to words in sentences is an

inverse function of the semantic congruency between the target word and preceding

context, even when the word is not semantically anomalous (Kutas & Hillyard, 1984).

N400 amplitude is also sensitive to the strength of semantic priming in a word-pair

lexical decision task (Bentin, McCarthy, & Wood, 1985). These results have been

interpreted as indicating that N400 amplitude elicited by words is highly sensitive to the

semantic relationship between the word and preceding context. However, N400

amplitude is not solely determined by the semantic fit between a word and its context. A

variety of lexical properties affect N400 amplitude, including lexicality (i.e., whether a

letter string is a word in the language), word frequency, phonological priming,

morphological content within a string (e.g., McKinnon, Allen, & Osterhout, 2003;

McLaughlin, Osterhout, & Kim, 2004), and sequential probabilities concerning the

likelihood of words occurring in succession (Kutas & Hillyard, 1984). A more accurate

generalization is that N400 amplitude reflects a combination of lexical and

semantic/conceptual factors.
 Osterhout et al.
8
The sensitivity of the N400 component to lexico-semantic factors is strikingly

contrasted with its insensitivity to syntactic factors. A great deal of research conducted

over the past 25 years has shown that violations of syntactic rules do not robustly affect

N400 amplitude (Osterhout & Nicol, 1999; Allen, Badecker, & Osterhout, 2003).

Instead, a disparate set of syntactic anomalies elicit a large positive wave in the ERP with

an onset at about 500 ms and a duration of several hundred ms (labeled the P600 effect

by Osterhout & Holcomb, 1992; Fig 1b). The P600 effect is highly sensitive to syntactic

variables but insensitive to lexico-semantic variables (Allen et al., 2003). Importantly, the

N400/P600 dichotomy is not a function of particular combinations of stimuli, task, and

language, but instead generalizes well across types of anomalies (including anomalies

involving phrase structure, morphosyntax, and constituent movement), numerous

stimulus and task conditions (e.g., rate of word presentation, subject’s task), modalities

(visual and auditory), and languages (including configurational languages like English

and non-configurational languages such as Japanese). In some reports, syntactic

anomalies have also elicited an anterior negativity within a window ranging from 150 to

500 ms, sometimes largest in amplitude over the left hemisphere (hence the label Left

Anterior Negativity, or LAN; Friederici, 1995; Neville, Nicol, Barss, Forster, & Garrett,

1991; Osterhout & Holcomb, 1992, 1993).

As is the case with any method, ERPs have their limitations. Although it is

relatively easy to determine the antecedent conditions (that is, the stimulus

manipulations) that produce or modulate some effect, it is usually very difficult to

identify the specific processes that are manifested by the effect. For example, even

though the P600 effect is reliably elicited by violations of a syntactic rule, it does not
 Osterhout et al.
9
necessary follow that the P600 effect itself reflects specifically syntactic (or even

linguistic) processes. This ambiguity exists because ERPs (just like neuroimaging

techniques such as fMRI) are correlational in nature. Consequently, the P600 effect

might reflect syntactic processes directly, or it might reflect some process that is highly

correlated with these syntactic processes, or with the processes that respond to a syntactic

anomaly. Deciding which of these possibilities is correct might prove to be an intractable

problem1. Fortunately, the correlations between manipulations of lexical and semantic

variables and changes in N400 amplitude, and between the presence of a syntactic

anomaly and the P600 effect, are very robust. One can therefore reasonably infer that a

linguistic anomaly was perceived to be semantically anomalous if it elicited an N400

effect, or syntactically anomalous if it elicited a P600 effect.

A second limitation is that ERPs usually need to be averaged twice, once for a

given subject and then over all subjects. This is required to achieve the necessary signal-

to-noise ratio. When the subjects are native speakers of a language, there seems to be

enough consistency (within and between subjects) in the ERP responses to linguistic

stimuli to produce robust and reproduceable results. However, a sample of second-

language learners will almost always be more variable than a sample of L1 speakers. It

seems quite likely that this variability will be expressed in the ERPs (as, for example,

greater variation in the latency or scalp distribution of particular ERP effects). If so, then

the differences in variability limit the types of inferences that are warranted by ERP data.

For example, if an effect is present in the L1 group but absent in the L2 group, this does

not necessarily entail that the process manifested by the effect is present in the L1 group

but not in the L2 group. This is because the effect might be present in both groups but
 Osterhout et al.
10
obscured in the L2 group due to larger variability in the effect’s timing and distribution.

The increased variability would reduce the size of the effect, or eliminate it altogether, in

the more variable population – even if it is present on individual trials, for individual

subjects. An example of this problem is provided by a study reported by Hahne (2001).

Hahne recorded ERPs to syntactic anomalies from native German speakers and from

Russian L1 speakers who had learned German after the age of 10. Both groups of

speakers showed a P600 effect to the anomalies, but only the native speakers also showed

a LAN effect. Hahne interpreted this result as indicating that the German learners were

missing the linguistic process manifested by the LAN effect. However, another

possibility is that the greater variability in the L2 group simply obscured the anterior

negativity, even if it was present on individual trials, in individual learners. This

possibility is likeliest for effects that are small in amplitude and have a limited temporal

extent (such as the LAN effect)2.

The increased variability associated with L2 learners complicates the use of ERPs

in another way. If the variability across subjects is large enough, the grand average

might not reflect the response shown by any particular individual learner, and might not

generalize to any identifiable population. That is, in such cases the “central tendency”

represented by significant effects in the grand average might reflect only the accidental

overlap of effects that were present in many learners in the sample, but might not reflect

larger and potentially more important effects that varied from learner to learner. If so,

then comparisons between, for example, L2 learners who were exposed to the L2 at

different ages might produce reliable differences between the groups but might not

represent true effects of age on acquisition. Fortunately, the validity or invalidity of such
 Osterhout et al.
11
results would become clear when the results prove difficult to replicate. Unfortunately,

this problem could easily lead to the publication of many unreplicable and irreconcilable

results.

ERPs therefore offer a great deal of promise for learning more about the

acquisition and use of a second language, but also carry with them certain risks. Our goal

has been to reap the benefits of ERPs while minimizing our exposure to the risks. To that

end, we have combined ERPs with longitudinal studies of novice language learners. Our

questions have included the following: How quickly is L2 knowledge about words and

sentences incorporated into the on-line, “real time” processing system? What variables

influence this rate? How quickly do learners’ brain responses to L2 words and sentences

begin to approximate their responses to L1 words and sentences? What is the relationship

between brain-based manifestations of learning and a person’s overt behavior? This

approach might also usefully investigate some of the standard questions within the field

of second-language research. For example, what influence does the learner’s L1 have on

acquisition of the L2? What role does a correspondence (or lack of correspondence)

between phonology and morphology play in the acquisition of a second language? Does

L2 learning progress in a continuous manner, or are there discrete stages of learning?

Although these questions are not directly about the cognitive neuroscience of L2, we

hope to show that the tools of cognitive neuroscience are particularly well-suited for

providing answers to them. And because ERPs directly reflect brain activity, the results

of these studies sometimes shed light on the cognitive neuroscience of second languages

even when that is not the explicit goal.

 Osterhout et al.
12
ERPs and L2 word learning

Research on bilingual lexical processing has primarily focused on the

organization of the bilingual lexicon (Kroll & Sunderman, 2003). Most theorists assume

that the representation of words is divided into two levels, a word form (lexical) level and

a word meaning (conceptual) level. The question of interest has been the degree of

independence of the two lexicons at these two levels (Kroll & Sunderman, 2003; Potter,

So, Von Eckhardt, & Feldman, 1984).

However, little is known about how L2 lexical and conceptual representations

develop during L2 acquisition (McLaughlin, 1998), and almost nothing is known about

the neurobiological correlates of this developmental process (for one recent exception,

see Raboyaeu, Marie, Balduyck, Gros, Demonet, & Cardebat, 2004). In a study

conducted in our laboratory, we examined word learning by measuring learning-related

changes to the N400 component of the ERP (McLaughlin, Osterhout, & Kim, 2004). As

noted above, N400 is sensitive to both lexical status (whether or not a particular wordlike

form is part of the language) and word meaning (Bentin, 1987; Kutas & Hillyard, 1980).

For native speakers of a given language, N400 amplitude is largest for pronounceable,

orthographically legal nonwords (pseudowords, e.g., flirth) intermediate for words

preceded by a semantically unrelated context, and smallest for words preceded by a

semantically related context. Our goal was to determine how much L2 exposure is

needed before the learner’s brain responses to L2 words and non-words resembles that of

a native speaker. Such results might reveal the rate at which French learners acquire

information about word forms and word meanings.

 Osterhout et al.
13
Our participants were English-speaking university students progressing through

their first year of classroom French instruction, and a control group who had never

received any French instruction. None of the students had significant exposure to French

prior to attending class. The stimuli were prime-target pairs of letter strings. Some of the

pairs contained two French words that were either semantically related (e.g., chien-chat

‘dog-cat’) or unrelated (maison-soif ‘house-thirst’). Other pairs contained a target that

was a pronounceable French pseudoword (mot-naisier). ERPs to the targets were of

interest. The French learners were tested three times in a longitudinal design: once near

the beginning of their French instruction (after ~14 hours on instruction), once near the

middle (~60 hours), and once near the end (~140 hours).We asked the French learners to

make word/nonword judgments to the words and pseudowords. We found that after only

14 hours of L2 instruction, the pseudowords elicited a robustly larger N400 than the

words (Fig. 2a, right column). This was true even though learners were at chance levels

(d’ = 0) when deciding if the letter strings were actual words in French. Furthermore, the

correlation between hours of instruction and the word/nonword N400 difference was very

robust (r = .72), suggesting that the N400 difference was approximately linearly related

to the learner’s exposure to the L2. This result suggests that the French learners rapidly

extracted enough information about French word forms so that their brains could

discriminate between actual words and pseudowords, even if the learners themselves

could not do so. Effects of word meaning, manifested as smaller N400s to words

preceded by a related than by unrelated word, were observed after ~60 hours of

instruction (Fig. 2b, right column). After ~140 hours of L2 instruction (Fig. 2c, right

column), the amplitude of the word/nonword and relatedness differences approximated

 Osterhout et al.
14
that typically observed with similar stimuli in native English or French speakers, even

though the learners’ explicit word/nonword judgments remained very poor (d’ < 1). No

differences in N400 amplitude across conditions were observed for a group of subjects

who had received no French instruction (Fig. 2a-c, left column).

These results show that the L2 learners extracted information about word form

after just ~14 hours of instruction, and information about word meaning by at least ~60

hours. Furthermore, these changes in brain activity immediately started to approximate

the responses seen in native speakers to analogous stimuli, and occurred even while the

learner’s conscious lexicality judgments were very poor. But what, exactly, were these

learners learning about words? The very early learning observed in this study might

involve elemental aspects of linguistic knowledge (e.g., knowledge about word forms)

that serve as prerequisites to L2 competence. People rapidly extract co-occurrence

statistics for letter and sound combinations within a language (Saffran, Johnson, & Aslin,

1999). Perhaps the word/nonword N400 effect reflected a similar type of learning. To

test that idea, we computed correlations between the word/nonword N400 effect and

bigram and trigram frequency (as computed from the relevant portions of the French

text). These correlations were very weak (p > .3). Another possibility is that learners

memorized whole-word forms they were exposed to during their L2 instruction. This

idea receives some support from the robust correlation between the N400 word/nonword

effect and the frequency of the target words in the learners’ French text (p = .67).

ERPs and L2 morphosyntactic learning.

 Osterhout et al.
15
A second line of research in our laboratory is motivated by the observation,

discussed above, that syntactic and semantic anomalies embedded within sentences elicit

distinct ERP effects (the P600 and N400 effects, respectively). This finding suggests that

separable syntactic and semantic processes exist. In an L2 learning context, one

implication is that L2 learners must somehow segregate linguistic input into those aspects

of the language that relate to sentence form and those that relate to sentence meaning.

That is, learners “grammaticalize” some aspects of the L2, but not others. Such a

conclusion introduces a host of interesting questions. How and when do L2 learners

grammaticalize aspects of the L2? Which aspects of the syntax are grammaticalized

first? What factors influence the rate and eventual success of grammaticalization? How

similar is the process of grammaticalization in learning an L1 or an L2? In our work,

what we mean by “grammaticalization” is specifically the instantiation of grammatical

knowledge into the learner’s on-line, real-time language processing system. Our

assumption is that, once a feature of the L2 has been grammaticalized, violations of that

aspect of the grammar should elicit a P600 effect.

To investigate such questions, we have focused on one particular aspect of

syntactic learning, namely, the acquisition of grammatical features and their associated

morphosyntactic rules. The centrality of grammatical features in recent formal theories

has led to renewed interest in studying the acquisition of the grammatical morphemes that

encode these features. Within linguistic theory (Radford, 1988), semantic properties of

language are encoded in word (lexical) categories such as Noun (N), Verb (V), and

Adjective (A); these categories are shared by all languages. Grammatical properties, by

contrast, are encoded in the morphology of grammatical categories such as noun phrase
 Osterhout et al.
16
(NP) and verb phrase (VP). Grammatical categories are associated with grammatical

features such as gender, number, and verbal person. These features (and how they are

involved in morphosyntax) vary across languages (Table 1). For example, with respect to

grammatical features, English has a number feature in the NP (e.g., boy vs. boys) whereas

French has both number and gender. With respect to morphosyntax, English and French

both have sentential agreement (i.e., agreement of the verb with the subject in verbal

person and number; e.g., I like vs. He likes), but only French has NP agreement, that is,

agreement between the noun and its determiner/ adjective in number and gender (e.g., le

garçon vs. les garçons; excluding the restricted English case of this/those) (Table 1).

Languages also differ with respect to the explicitness of grammatical features. For

example, although English and French both show agreement between subject and verb,

the explicitness of the grammatical marking differs. English present tense uses only –s in

third person singular (he walks, they/I/etc. walk). French present uses five orthographic

forms (je marche, tu marches “I/you walk”), but has only three distinct pronunciations.

As for NP agreement, English marks number on the noun alone (the little boy[s]),

whereas French marks number (and gender) on the agreeing determiner/adjective (le[s]

petit[s] garçon[s) (Table 1). In oral French, however, plurality can only be heard on the

determiner. English has no grammatical gender feature whereas French does.

What factors might inhibit or facilitate grammaticalization of these features and

their morphosyntactic rules? One frequent claim is that only features that are present in

the L1 can be acquired during L2 acquisition (Franceschina 2001, Hawkins &

Franceschina 2004). Other researchers, however, argue that novel L2 features can be

learned, albeit more slowly than those that are present in the L1 (White, 2003). Thus,
 Osterhout et al.
17
there is no consensus about whether, or when during acquisition, L2 learners acquire L2

features that are not present in their L1. Even less is known about the acquisition of L2

morphosyntactic rules (e.g., agreement in number or gender) that are not present in the

L1.

Another factor that seems likely to play a role in L2 grammatical morpheme

learning is the covariation between morphology and phonology. The interaction between

morphology and phonology can be clearly seen in written French. French has an opaque

orthography due to many suffixes being phonologically silent (Table 1). Thus, the plural

suffix –s, which marks the plural across all elements in the NP (le-s jeune-s fille-s “the

young girls”) is silent on the noun (as well as on the adjective) in almost all instances. A

similar situation arises in the VP, where variations in verbal person are marked

orthographically on the verb but are silent in most oral forms. Thus the different

inflections for a regular verb such as marcher (to walk) sound identical across three

different persons/spellings (Table 1). The effect of the ‘missing’ phonological cue is

notorious on spelling. French children are drilled in the morphological variations

throughout elementary school, and yet they continue to make grammatical errors, by

failing to add the plural inflection in writing. Errors such as les chien or Ils mange are

frequent (Negro & Chanquoy, 2000), and can be seen in adults as well. Errors are much

rarer when phonology is available as a cue (Largy & Fayol, 2001). For example, in the

VP, confusions are not made among persons for nous and vous (we/you formal or plural),

as these forms are both morphologically and phonologically distinct from each other and

all other persons. Errors are also far less frequent when the morphological ending is

paired with a phonological difference, as is the case for certain irregular and stem-
 Osterhout et al.
18
changing verbs in French. The results from French are in line with results from studies of

Dutch spelling (Frisson & Sandra, 2002; Sandra, Frisson, & Daems, 1999) showing that

adult Dutch writers continue to confuse the first and third person singular for verbs where

these two persons have distinct morphological endings but identical oral realizations (e.g.

rijd “drive” and rijdt “drives”, which are both pronounced as /r It/), but do not confuse

the two for verbs where they have distinct phonologies (e.g. werkt /w rkt/ “works” and

werk /w rk/ “works”). In adult L2 acquisition, the use of oral cues has been advocated by

linguists to enhance the learning of certain morphological rules, such as verb and

adjective variations (Herschensohn, 1993). Arteaga, Gess, and Herschensohn (2003)

found that English learners of French who were instructed to note phonological variations

of masculine and feminine forms of variable adjectives in French were quicker to learn

the morphological rule than were students who were not given such instruction, but who

were taught only the spelling alterations between the two forms.

Therefore, one reasonable prediction is that L2 learners will acquire an L2 feature

or morphosyntactic rule more quickly when the relevant inflectional morphology is

phonologically realized. However, this possibility has received little direct attention in

the recent L2 literature. It also seems likely that L1-L2 similarity and phonological-

morphological covariation might have interactive effects during L2 learning. For

example, L1-L2 similarity combined with phonological realization of the relevant

grammatical morphemes might lead to very fast learning, whereas L1-L2 dissimilarity

combined with no phonological realization might lead to very slow learning.

Finally, various theorists have suggested that L2 morphosyntactic learning

involves discontinuous stages (Myles, Hooper, & Mitchell, 1998; Wray, 2002; Wong-
 Osterhout et al.
19
Fillmore, 1976; Vainikka and Young-Scholten, 1996). This seems to be true of children

learning their first language, who begin by memorizing particular combinations of words

and only later induce general syntactic rules (Tomasello, 2000). With respect to L2

leaning, one account holds that morphologically complex words are initially learned by

rote and memorized as unanalyzed chunks (Myles et al., 1998; Wray, 2002). Eventually,

learners unpack these chunks into roots and grammatical morphemes, induce

grammatical rules governing their use, and deductively use these units to produce novel

utterances. To be specific, an L2 learner might initially memorize the fact that certain

subjects are followed by certain forms of the verb, without decomposing the verb into

root + inflection or applying a general morphosyntactic agreement rule. In this stage of

learning, the learner associates meanings with the undecomposed word, and either

memorizes the two words as a chunk or learns about word sequence probabilities (e.g.,

that Tu is followed by marches, whereas Il is followed by marchent)2. Unfortunately,

using conventional behavioral methods it is very difficult to identify the transition from

rote-memorization to rule use (Wray, 2002). Fortunately, ERPs might be an ideal tool for

testing this theoretical claim of developmental discontinuity. As noted above, N400

amplitude is highly sensitive to novel words and to word sequence probabilities, whereas

the P600 effect is sensitive to grammatical rule violation. If an L2 learner initially

memorizes salient word sequences, then unfamiliar word combinations (e.g., *Tu adorez)

should produce larger N400s than familiar ones (e.g., Tu adores). If the learner

eventually decomposes the verb into root + morpheme and induces the rule, use of the

wrong verbal form should elicit a P600 effect. In essence, the transition for N400 to
 Osterhout et al.
20
P600 would reflect the transition from rote-memorized (i.e., lexicalized) knowledge to

grammaticalized knowledge.

We investigated these predictions using a longitudinal experimental design

involving 14 English-speaking novice French learners. Each learner was tested after

approximately one month, four months, and eight months of university classroom French

instruction. Our stimuli were as follows:

(1) Sept plus cinq\?livre font douze.

semantic condition

(2) Tu adores\*adorez le francais.

verbal person agreement condition/phonologically realized

(3) Tu manges des hamburgers\*hamburger pour diner.

number agreement condition/phonologically unrealized

In (1), the noun livre is semantically anomalous. In (2), the verb adorez is conjugated

incorrectly, given the preceding sentence fragment. In (3), the noun hamburger disagrees

with the syntactic number of the plural article. Our stimuli were selected from the

material in the textbook assigned during the first month of instruction. The anomalous

items in the verbal person condition involved a grammatical rule that was present in the

L1 and an orally realized contrast between inflectional morphemes. The anomalous

items in the number agreement condition involved a rule that was not present in the L1

and a phonologically unrealized contrast between inflectional morphemes. Therefore,

 Osterhout et al.
21
our prediction was that L2 learners would respond to the anomaly in (2) with less L2

exposure compared to the anomaly in (3).

As expected, the native French speakers showed an N400 effect to the

semantically anomalous words and large P600 effects to the two types of syntactic

anomalies. The learners, as is often the case, showed striking individual differences, both

in a behavioral “sentence acceptability judgment” task and in the pattern of ERPs elicited

by the anomalous stimuli. We segregated the learners into upper (“fast learners”) and

lower (“slow learners”) halves, based on their performance in the sentence-acceptability

judgment task that occurred concurrently with ERP data collection, and averaged the

ERPs separately for each group. Results for the “fast learner” group will be described

here. At each testing session, including the initial session that occurred after just one

month of instruction, semantically anomalous words elicited a robust N400 effect, and

this effect changed minimally with increasing instruction (Fig. 3, averaged over the three

testing session.). Results for the verbal person condition are shown in Fig. 4. After just

one month of instruction, the learners’ brains discriminated between the syntactically

well-formed and ill-formed sentences. However, rather than eliciting the P600 effect (as

we saw in native French speakers), the syntactically anomalous words elicited an N400-

like effect. (This effect did not differ in distribution from the N400 effect elicited by the

semantically anomalous words.) By four months, the N400 effect was replaced by a

P600-like positivity. Results for the number agreement condition can be summarized

easily: Learners performed very poorly in the sentence acceptability judgment task for

these materials, and there were no robust differences between the agreeing and

disagreeing stimuli (Fig. 5, averaged over 3 testing sessions).

 Osterhout et al.
22
These results are consistent with the predictions we were testing. First, L1-L2

similarity combined with phonological realization of the relevant grammatical

morphemes produced very fast L2 syntactic learning, whereas L1-L2 dissimilarity

combined with no phonological realization produced very slow learning. This occurred

even though our learners were drilled repeatedly on both rules from nearly the first day in

class. However, the two rules we tested represent the ends of a putative continuum of

morphosyntactic difficulty; without additional data it is impossible to know whether L1-

L2 similarity or phonological realization of grammatical morphemes had a larger impact

on the learning rate. What can be said with more certainty is that at least some times of

L2 syntactic rules are learned amazingly quickly. Apparently, at least some types of L2

rules are incorporated into the on-line sentence processing system after a very small

amount of L2 instruction.

The second prediction was that L2 morphosyntactic anomalies would show a

discontinuous pattern over time: early in learning, such anomalies would elicit an N400

effect in learners, whereas later in learning these same anomalies would elicit a P600

effect. This pattern is exactly what we observed, for the verbal person anomalies. If our

interpretation is correct, then our adult L2 learners grammaticalized this aspect of the L2

after just a few months of L2 instruction. Our results can be explained by assuming that

learners (much like child L1 learners) initially memorize salient word sequences (e.g., tu

adores). Violations of the verbal person rule (e.g., tu adorez) result in novel word

combinations, and hence elicit an N400 effect. After more instruction, learners induce a

general verbal person rule (tu -s, nous –ons, vous -ez, etc); violations of the rule elicit a

P600 effect. The P600 effect occurred at similar temporal latencies for the natives and
 Osterhout et al.
23
learners. Thus, after just 80 h of instruction, learners’ and natives’ ERP responses to

verbal person violations were qualitatively similar.

English and French are highly similar in many respects; for example, they share

many cognates and a similar (although not identical) system of morphosyntax, and they

both belong to the Indo-European family. English and Finnish, by contrast, share very

few cognates and come from different language families. English has a relatively

impoverished morphosyntax, whereas Finnish has an extensive system: nouns are marked

for number, case, and possessiveness, and finite verbs are marked for tense or mood, and

number and person. Agreement rules in Finnish exist for number, person, and case.

Finnish is generally thought to have a subject-verb-object canonical structure. However,

since the morphosyntactic system usually makes the syntactic and semantic functions of

each constituent in a clause unambiguous, Finnish speakers have considerable flexibility

in ordering words. Noncanonical structures are permitted to mark a change in focus or for

other purposes. Furthermore, the Finnish morphosyntactic system is highly regular, and

fully realized in the phonology. Inflectional suffixes are regular, and they are added to the

word stems in a fixed order. However, the addition of the endings is sometimes

accompanied by sound alternations in the stem of the word.

These properties of Finnish morphosyntax provide a useful contrast with French,

in the context of a native English-speaking L2 learner. If L1-L2 similarity is the primary

determinant of the rate of L2 syntactic learning, then learners should be quite slow to

acquire the many novel aspects of Finnish morphosyntax. If, conversely, regularity,

transparency, and phonological realization in the morphosyntax are more important, then

learners might acquire the system rather quickly.

 Osterhout et al.
24
Recent work has shown that morphosyntactic and semantic anomalies elicit the

P600 and N400 effects, respectively, when the subjects are native speakers of Finnish

(Palolahti, Leino, Jokela, Kopra, & Paavilainen, 2005). In our lab, we have begun to

study English-speaking university students as they progress through their first year of

Finnish instruction. We hope to determine how much Finnish instruction is needed

before morphosyntactic violations elicit an ERP linguistic anomaly effect. We are

presenting sentences that contain either a semantic anomaly and one of two kinds of

morphosyntactic errors, as illustrated below:

(4a) Sinä (2nd p. sing.) syöt (2nd p. sing.) lihaa.

(‘You eat meat.’)

(4b) *Sinä (2nd p. sing.) syön (1st p. sing.) lihaa.

(*’You eat meat.’)

(5a) He istuvat kahvilassa (inessive case) illalla.

(‘They sit in a café at night.’)

(5b) *He istuvat kahvilasta (elative case) illalla.

(* ‘They sit from a café at night.’)

In sentence (4b), the verb is in the wrong person form. In sentence (5b), the noun is

attached to an incorrect case particle. Because English has a verbal person rule but not a

system of explicit case marking, learning might be relatively fast for Finnish verbal

person but very slow for case, if L1-L2 similarity is the primary determinant of learning
 Osterhout et al.
25
rate. Conversely, if phonological realization of the grammatical morphemes is very

important, then both the similar and dissimilar aspects of the morphosyntactic system

might be learned relatively quickly.

In a pilot study, we tested five English speakers who had just completed their

initial nine months of Finnish instruction. The results were as follows: Semantic

anomalies elicited an N400 effect; verbal person anomalies elicited a robust P600 effect;

and case anomalies also elicited a P600-like effect, although with a much smaller

amplitude. These preliminary data suggest that even unfamiliar aspects of a complex

morphosyntactic system can be grammaticalized within the first year of classroom

instruction. Our ongoing research with this population might tell us more about the rate

of learning, and whether this type of morphosyntactic learning is continuous or

discontinuous.

Conclusions and final comments

Our goal in this chapter has been to describe a research paradigm for investigating

what L2 learners learn as they progress through a period of L2 instruction, and when they

learn it. By learning, we mean specifically the incorporation of L2 knowledge into the

learner’s on-line, real-time language processing system. We do not claim to have

definitive answers to any of the questions we hope to answer. For example, our pilot

Finnish study suggests that English-speaking L2 learners might “grammaticalize” novel

aspects of Finnish morphosyntax during their initial 8 months of classroom instruction.

We have suggested that this might be due, in part, to the fact that the relevant
 Osterhout et al.
26
grammatical morphemes are reliably (and uniquely) expressed in the language’s

phonology. But of course there are many other explanations for this effect, if it turns out

to be robust. Our proposed paradigm is limited in other ways as well. Although the

paradigm minimizes between-subjects variability, it cannot eliminate it. As we found in

our French syntax learning study, individual differences continue to exert an influence on

learning even when we longitudinally study novice learners in a homogeneous classroom

environment. This is not unexpected. Our hope is that our design reduces between-

learner variability to a manageable level; eventually, careful study might reveal the

primary sources of the remaining variability. Finally, it is important to explicitly

recognize that ERPs (like all methods of investigation) imperfectly reflect the cognitive

and neural processes underlying L2 learning and processing.

Even given these caveats, our preliminary findings are promising. Our findings

suggest that it is possible to ascertain when some aspect of the L2 has been incorporated

into the set of processes that allow the learner to comprehend the language in real time.

Moreover, our results suggest that developmental discontinuities exist with respect to

what has been learned: early in the learning process, learners seem to memorize

frequently occurring words or groups of morphemes. With a little more L2 exposure,

learners associate the memorized lexical forms with meanings, begin to decompose the

structurally complex units into roots plus grammatical morphemes, and learn the rules for

agreement among these elements. This type of learning can occur with remarkably little

L2 exposure, in a standard classroom setting that is often thought to provide a less-than-

ideal learning environment. Some learners even seem to quickly grammaticalize novel

aspects of a complex L2 morphosyntactic system that is unlike anything in their L1. Our
 Osterhout et al.
27
results are preliminary, but they clearly conflict with the conventional belief that adult L2

syntactic learning is generally slow and problematic, especially for aspects of the L2 that

are not present in the L1. The conventional belief about the difficulty of L2 syntactic

learning is based, in part, on L2 production: many speakers do have difficulty with

grammatical aspects of a second language, when they try to speak it. However, one

general rule about L1 learning is that a learner’s ability to understand the language

develops in advance of her ability to produce it. It seems likely that this maxim also

applies to L2 learners. If so, then perhaps there is no contradiction between our data

(indicating a rapid acquisition and implementation of some types of L2 knowledge during

language comprehension) and the common impression that L2 learners have a difficult

time producing certain aspects of the L2.

We should note that other researchers have come part way in implementing our

proposed research paradigm, although (to the best of our knowledge) no one has

implemented it completely. Several groups of researchers have used ERPs to study the

incorporation of L2 knowledge into the on-line language processing system (e.g., Hahne,

2001). However, the learners in these studies were not novice learners but instead had

extensive (and possibly highly variable) experience with their L2. Furthermore, none of

these studies provided a contrast between different levels of L2 exposure or proficiency.

In other studies, researchers have recorded ERPs from subjects who were acquiring

aspects of an artificial language (Friederici et al., 2002; McCandliss, Posner, & Givon,

1997). As in our studies described above, the learners were novices with respect to the

“language” to be learned. Although these studies are quite interesting, it is unclear to

what degree the results can be generalized to the acquisition of a natural language in a
 Osterhout et al.
28
typical (within the United States) L2 learning environment. These studies have also not

included a contrast involving learners at different stages of acquisition.

Only a few studies have both recorded ERPs and contrasted different stages of L2

exposure or proficiency. For example, Weber-Fox and Neville (1996) examined the ERP

response to linguistic anomalies in adult Chinese-English bilinguals who had been

exposed to English at various ages, ranging from 1 to 16 years. They report interesting

differences in the ERP responses to a variety of linguistic anomalies. However, because

the study was cross-sectional in nature, there is inevitably some ambiguity about whether

these different ERP patterns are in fact manifestations of age of exposure effects, or

instead reflect some other source of uncontrolled subject variability.

The approach proposed here is not a panacea and is itself open to criticism. For

example, at the moment we cannot predict or explain learner differences of the type we

observed in our French syntax learning study. It is also true that longitudinal designs are

extremely time-consuming and are not always viable. Nonetheless, this research

paradigm might produce novel results that are less subject to the ambiguity plaguing

much of the existing literature. By focusing on changes in the neural manifestations of

L2 processing that occur with increasing L2 instruction within a set of L2 learners, we

hope to demonstrate the value of asking the what and when questions: Exactly what

processes change over time with increasing exposure to or proficiency in a second

language? And when (in the course of understanding a word or sentence, and in the

course of learning a language) do those changes occur? Some progress has already been

made toward answering such questions in the domain of L2 speech perception (Kuhl,
 Osterhout et al.
29
2004). Perhaps we can reasonably hope that similar advances will soon be forthcoming

with respect to the on-line processing of L2 words and sentences.

 Osterhout et al.
30
Acknowledgements

The authors thank Julia Herschensohn for her comments and collaboration. We received

financial support from Grants R01DC01947 and P30DC04661 from the National Institute

on Deafness and Other Communication Disorders, National Institutes of Health.

 Osterhout et al.
31
References

Ainsworth-Darnell, K., Shulman, R., & Boland, J. (1998). Dissociating brain responses to

syntactic and semantic anomalies: Evidence from event-related brain potentials. Journal

of Memory and Language, 38, 112-130.

Allen, M. D., Badecker, W., & Osterhout, L. (2003). Morphological analysis during sentence

processing. Language and Cognitive Processes, 18, 405-430.

Arteaga, D., Gess, R., & Herschensohn, J. (2003). Focusing on phonology to teach

morphological form in French. Modern Language Journal, 8, 58-70.

Bentin, S. (1987). Event-related potentials, semantic processes, and expectancy factors in word

recognition. Brain and Language, 31, 308-327.

Bentin, S., McCarthy, G., & Wood, C. (1985). Event-related potentials, lexical decision, and

semantic priming. Electroencephalography and Clinical Neurophysiology, 60, 343-355.

Chee, M. W. L., Tan, E. W. L., & Thiel, T. (1999). Mandarin and English single word processing

studies with functional magnetic resonance imaging. Journal of Neuroscience, 19, 3050-

3056.

Clahsen, H., Penke, M., & Parodi, T. (1994). Functional categories in early child German.

Language Acquisition, 3, 395-429.

Franceschina, F. (2001). Morphological or syntactic deficits in near–native speakers? An

assessment of some current proposals. Second Language Research, 17, 213–247.

Friederici, A. D. (1995). The time course of syntactic activation during language processing: a

model based on neuropsychological and neurophysiological data. Brain and Language,

50, 259-284.
 Osterhout et al.
32
Friederici, A. D., Steinhauer, K., & Pfeifer, E. (2002). Brain signatures of artificial languages:

Evidence challenging the critical period hypothesis. Proceedings of the National

Academy of Sciences, 99, 529-534.

Frisson, S. & Sandra, D. (2002). Homophonic forms of regularly inflected verbs have their own

orthographic representations: A developmental perspective on spelling errors. Brain and

Language, 81, 545-554.

Grosjean, F. (1998). Studying bilinguals: Methodological and conceptual issues. Bilingualism:

Language and Cognition, 1, 131-149.

Hagoort, P., Brown, C., & Groothusen, J. (1993). The syntactic positive shoft (SPS) as an ERP

measure of syntactic processing. Language and Cognitive Processes, 8, 439-483.

Hahne, A. (2001). What’s different in second-language processing? Evidence from event-related

brain potentials. Journal of Psycholinguistic Research, 30, 251-266.

Hahne, A., & Fiederici., A. D. (1999). Electrophysiological evidence for two steps in syntactic

analysis: early automatic and late controlled processes. Journal of Cognitive

Neuroscience, 11, 193-204.

Hahne, A. & Friederici, A. D. (2001). Processing a second language: Late learners

comprehension mechanisms as revealed by event-related brain potentials. Bilingualism:

Language and Cognition, 4, 123-141.

Hawkins, R. & Franceschina, F. (2004). Explaining the acquisition and non-acquisition of

determiner-noun gender concord in French and Spanish. In P. Prevost & Johanne Paradis

(Eds.), The acquisition of French in different contexts. Amsterdam: Benjamins.

Herschensohn, J. (1993). Applying linguistics to teach morphology: Verb and adjective

inflection in French. International Review of Applied Linguistics 30, 97-112.

 Osterhout et al.
33
Kim, A. & Osterhout, L. (2005). The independence of combinatory semantic processing:

Evidence from event-related potentials. Journal of Memory and Language, 52, 205-225.

Kim, K. H. S., Relkin, N.R., Lee, K-M, & Hirsch, J. (1997). Distinct cortical areas associated

with native and second languages. Nature, 388,171-174.

Klein, D., Milner, B., Zatorre, R., Zhao, V., & Nikelski, J. (1999). Cerebral organization in

bilinguals: Chinese-English verb generation. NeuroReport, 10, 2841-2846.

Kroll, J. & Sunderman, G. (2003). Development of lexical and conceptual representations. In C.

Doughty & M. Long (Eds), The handbook of second-language acquisition. Malden, MA:

Blackwell

Kuhl, P. (2004). Early language acquisition: Cracking the speech code. Nature Rewviews

Neuroscience, 5, 831-843.

Kuperberg, G., Holcomb, P. J., Sitnikova, T., Greve, D., Dale, A. M., & Caplan, D. (2003).

Distinct patterns of neural modulation during the processing of conceptual and syntactic

anomalies. Journal of Cognitive Neuroscience, 15, 272-293.

Kutas, M., and Hillyard, S. A. (1980). Reading senseless sentences: Brain potentials reflect

semantic anomaly. Science, 207, 203-205.

Kutas, M., & Hillyard, S. A. (1984). Brain potentials during reading reflect word expectancy and

semantic association. Nature, 307, 161-163.

Largy, P. & Fayol, M. (2001). Oral cues improve subject-verb agreement in written French.

International Journal of Psychology, 36, 121-131.

McCandliss, B.D., Posner, M.I., & Givon, T. (1997). Brain plasticity in learning visual words.

Cognitive Psychology, 33, 88-110.

 Osterhout et al.
34
McKinnon, R., Allen, M., & Osterhout, L. (2003). Morphological decomposition involving non-

productive morphemes: ERP Evidence. NeuroReport, 14, 883-886.

McKinnon, R., & Osterhout, L. (1996). Constraints on movement phenomena in sentence

processing: Evidence from event-related brain potentials. Language and Cognitive

Processes, 11, 495-523.

McLaughlin, B. (1998). Second language learning revisited: The psycholinguistic perspective. In

A. F. Healy and L. E. Bourne, Jr. (Eds.), Foreign Language Learning: Psycholinguistic

Studies on Training and Retention. Mahwah, NJ: Erlbaum.

McLaughlin, J., Osterhout, L., & Kim, A. (2004). Neural correlates of second-language word

learning: minimal instruction produces rapid change. Nature Neuroscience, 7, 703-704.

Myles, F., Hooper, J. & Mitchell, R. (1998). Rote or Rule? Exploring the Role of Formulaic

Language in Classroom Foreign Language Learning. Language Learning, 48, 323-63.

Negro, I., & Chanquoy, L. (2000). Subject-verb agreement with present and imperfect tenses: a

developmental study from 2nd to 7th grade. European Journal of Psychology of

Education, 15, 113-134.

Neville, H. J., Nicol, J. L., Barss, A., Forster, K. I., & Garret, M.(1991). Syntactically based

sentence processing classes: Evidence from event-related brain potentials. Journal of

Cognitive Neuroscience, 3, 151-165.

Osterhout, L., Bersick, M., & McLaughlin, J. (1997). Brain potentials reflect violations of gender

stereotypes. Memory and Cognition, 25, 273-285.

Osterhout, L., & Holcomb, P. J. (1992). Event-related brain potentials elicited by syntactic

anomaly. Journal of Memory and Language, 31, 785-806.

 Osterhout et al.
35
Osterhout, L., & Holcomb, P. J. (1993). Event-related potentials and syntactic anomaly:

Evidence of anomaly detection during the perception of continuous speech. Language

and Cognitive Processes, 8, 413-438

Osterhout, L., McLaughlin, J., Kim, A., & Inoue, K. (2004). Sentences in the brain: Event-

related potentials as real-time reflections of sentence comprehension and language

learning. In M. Carreiras & C. Clifton, Jr. (eds.), The on-line study of sentence

comprehension. Psychology Press.

Osterhout, L., & Mobley, L. A. (1995). Event-related brain potentials elicited by failure to agree.

Journal of Memory and Language, 34, 739-773.

Osterhout, L., & Nicol, J. (1999). On the distinctiveness, independence, and time course of the

brain responses to syntactic and semantic anomalies. Language and Cognitive Processes,

14, 283-317.

Palolahti, M., Leino, S., Jokela, M., Kopra, M., & Paavilainen, P. (2005). Event-related

potentials suggest early interaction between syntax and semantics during on-line sentence

comprehension. Neuroscience Letters, 384, 222-227.

Perani, D. Paulesu, E., Galles, N.S., Dupoux, E. Dehaene, S., Bettinardi, V., Cappa, S., Fazio, F.,

& Mehler, J. (1998). The bilingual brain: Proficiency and age of acquisition of the second

language. Brain, 121, 1841-1852.

Potter, M. C., So, K. – F., Von Eckardt, B., & Feldman, L. B. (1984). Lexical and conceptual

representation in beginning and more proficient bilinguals. Journal of Verbal Learning

and Verbal Behavior, 23, 23-38.

 Osterhout et al.
36
Raboyeau, G., Marie, N., Balduyck, S., Gros, H., Demonet, J., & Cardebat, D. (2004). Lexical

learning of the English language : a PET study in healthy French subjects. NeuroImage,

22, 1808-1818.

Radford, A. (1988). Transformational grammar. Cambridge, UK: Cambridge University Press.

Saffran, J.R., Johnson, E.K., Aslin, R.N., & Newport, E.L. (1999). Statistical learning of tonal

sequences by human infants and adults. Cognition, 70, 27-52.

Sandra, D., Frisson, S., & Daems, F. (1999). Why simple verb forms can be so difficult to spell:

The influence of homophone frequency and distance in Dutch. Brain and Language, 68,

277-283.

Tomasello, M. (2000). Trends in Cognitive Sciences, 4, 156-160.

Vainikka, A., & Young–Scholten, M. (1996) Gradual development of L2 phrase structure.

Second Language Research, 12, 7–39.

Weber-Fox, C. & Neville, H. J. (1996). Maturational constraints on functional specializations for

language processing: ERP and behavioral evidence in bilingual speakers. Journal of

Cognitive Neuroscience, 8, 231-257.

White, L. (2003). Second Language Acquisition and Universal Grammar. Cambridge:

Cambridge University Press.

Wong-Fillmore, L. (1976). The second time around: Cognitive and social strategies in second

language acquisition. Unpublished doctoral dissertation, Stanford University, Stanford,

CA.

Wray, A. (2002). Formulaic language. Cambridge: Cambridge University Press.

 Osterhout et al.
37
Footnotes
1
Friederici and colleagues (Hahne & Friederici, 1999; Friederici, 1995) have proposed a

two-stage functional model for these ERP effects. They claim that the anterior negativity

reflects a fast, automatic syntactic analyzer, and the P600 reflects syntactic reanalysis. As

we have explained elsewhere (Osterhout et al., 2004, pp 294-298), we do not believe that

the available evidence strongly supports these theoretical claims.

2
Another problematic aspect of using the LAN effect to study L2 processing is that there

are a significant number of reports involving native speakers in which they are not

reported (e.g., Ainsworth-Darnell, Shulman, & Boland, 1998; Allen, Badecker, &

Osterhout, 2003; Hagoort, Brown, & Groothusen, 1993; Kuperberg, Holcomb, et al.,

2003; Kim & Osterhout, 2005; McKinnon & Osterhout, 1996; Osterhout, Bersick, &

McLaughlin, 1997; Osterhout & Mobley, 1995). This makes it difficult to interpret the

absence of the effect in L2 learners.

3
A related account of L2 learning (Vainikka and Young-Scholten, 1996) proposes that

L2 learners initially transfer their L1 lexical categories, but not their L1 grammatical

features and categories; learners are claimed to infer these features and categories from

the presence of the associated L2 grammatical morphemes. Thus, syntactic structure is

projected from the acquired lexical material, and learners must go through a stage of

lexical learning before they acquire the features and associated morphosyntax (see also

Clahsen, Penke, & Parodi, 1994).

 Osterhout et al.
38
Table 1

Sentential verb-subject agreement: Regular Verbs,

present tense

English French

Singular
1st I like Je marche
2nd You like Tu marche(s)
3rd He likes Il marche

Plural
1st We like Nous marchons
2nd You like Vous marchez
3rd They like Ils marche(nt)

Noun phrase (NP) number agreement

English French

Singular the boy le petit garçon

Plural the boys les petit(s) garçon(s)

Noun phrase (NP) gender agreement

English French
Male NA le garçon
Female NA la fille

Note: morphemes in parentheses are not expressed

phonologically
 Osterhout et al.
39
Figure Captions

Fig.1. (a) ERPs (recorded at central midline location Cz) elicited by semantically

anomalous words (dashed line) and nonanomalous control words (solid line) in sentences

such as The cat will eat/bake the food. (b) ERPs elicited by syntactically anomalous

(dashed line) words and well-formed controls (solid line) in sentences such as The cat

will eat/eating the food. Onset of the critical word is indicated by the vertical bar. Each

hashmark represents 100 ms. The vertical calibration bar is 5 µV. Adapted from

Osterhout and Nicol (1999).

Fig. 2. ERPs from the No Instruction (left panel) and French Instruction (right panel)

subjects, recorded at three successive longitudinal testing sessions. ERPs are plotted for

three types of target strings: target words that were semantically related to the prime word

(solid line), target words that were not semantically related to the prime (small dashes),

and pseudoword targets (large dashes). Adapted from McLaughlin, Osterhout, and Kim

(2004).

Fig 3. ERPs to critical words in the well-formed (solid line) and semantically anomalous

(dashed line) conditions, collapsed over the 3 testing sessions.

Fig. 4. ERPs to critical words in the well-formed (solid line) and verbal person anomaly

(dashed line) conditions, plotted separately for each of the three testing sessions.

Fig. 5. ERPs to critical words in the well-formed (solid line) and noun phrase agreement

anomaly (dashed line) conditions, collapsing over the three testing sessions.
 Osterhout et al.
40

Fig. 1

a N400

The cat will EAT

The cat will BAKE

P600

The cat will EAT

*The cat will EATING
 Osterhout et al.
41

Fig. 2

No French Instruction French Learners

N400

Session 1

Session 2

Session 3
 Osterhout et al.
42
Fig. 3

Well-formed

Semantically anomalous
 Osterhout et al.
43
Fig. 4

Session 1 Session 2 Session 3

_____ Well-formed

- - - - - Verbal person anomaly

Well-formed

Verbal person anomaly

 Osterhout et al.
44
Fig. 5

Well-formed

Noun phrase agreement anomaly