16

Input processing

Bill VanPatten

Historical discussion

Input processing emerged as a field of inquiry in the mid-1980s. The central question it sought to address was “What linguistic data do learners process in the input and what constrains/guides that processing?” As such, research on input processing was a natural outgrowth of the importance given to input within the field of second language acquisition (SLA) during the 1970s and 1980s. At that time, research tended to focus on how input was negotiated by learners and/or modified by other speakers (e.g., Hatch, 1983; Long, 1983; and the many papers in Gass and Madden, 1985). The central concern of the time was what made input comprehensible and thus useable by second language (L2) learners. An underlying assumption of the research seemed to be that if input was comprehensible, learners would comprehend it correctly and get good linguistic data from that input (see, for example, Krashen, 1982).

But it became increasingly clear that although negotiation was probably important to SLA more generally, it was insufficient as a construct to account for acquisition. Something internal to the learner was interacting with input in particular ways, as some kind of filter or set of strategies that kept input from being completely useful. As I asked in the mid-1980s, why don't learners seem to pick up things that they hear in another person's question or statement, particularly when that statement is relatively short and should not tax working memory limitations (VanPatten, 1984a)? The reader is invited to examine the following interchange (E = experimenter, L = learner; all errors are original to learner):

E: ¿Cómo están ellos? “How are they?”

L: Ellos son contentos, “They're happy.”

E: Y ellos, ¿cómo están?, “And them. How are they?”

L: Son contento también. “They're happy, too.”

( from VanPatten, 1984a, p. 92)

Only estar can be used as a copular verb with predicate adjectives possessing the feature [+perfective], whereas ser is used with predicates that are [-perfective] in nature. In the above interchange, the experimenter uses the correct copular verb (estar) in Spanish to express a condition [+perfective] but the learner uses the incorrect copular (ser). Note that in his second question, the experimenter follows standard conventions for emphasis and change of topic and moves the copular to the end of the sentence, that is, ¿cómo están? This second question is not long and there is no new information that the learner has to attend to: both speakers used the same words throughout the interchange. What is more, in the E's second question the copular verb now has two things going for it in terms of perceptual salience: it is in final position and it carries strong stress. Why can't the learner pick up the copular verb during this interchange? Is the learner even paying attention to it? If not, why not? As this example illustrates, exposure to a linguistic item in the input does not mean that item gets acquired at that time.

Work on child first language (L1) acquisition was also suggestive that learners somehow filter input (e.g., Bever, 1970; Slobin, 1966). For example, children learning their L1s seemed to misinterpret passive structures (“The lion was killed by the tiger”) as active structures (“The lion killed the tiger”). What this suggested was that during acquisition, even though the input to children was “comprehensible,” they didn't comprehend it correctly. In 1974, Ervin-Tripp demonstrated with child L2 learners that even if literal translations of passives existed in two languages (in her study, French and English) children who already have passives in their L1 do not use the syntactic information they possess to interpret passives early in their acquisition of the L2. Just like L1 children, they would misinterpret passive sentences as active sentences. Thus, even though English and French have word for word passive equivalents, her learners interpreted the L2 passives as actives, ignoring all the passive cues in the sentences.

The point here is that in the1970s and 1980s, L2 researchers were looking at the role of input globally in terms of input modification and negotiation and not psycholinguistically in terms of processing. They were not asking the more micro-level question of how learners comprehended individual sentences and what elements of sentences learners attended to. Input processing as an area of research emerged to address this gap in the field.

To be sure, input processing has never been viewed as a model of acquisition per se, but as part of a larger acquisition puzzle. Thus, it is compatible with a variety of other approaches and frameworks. One way to conceptualize this is to consider that input processing research focuses on one and only one set of processes used by learners during the course of acquisition. There are other processes and factors that must be involved in order to have a full understanding of SLA (e.g., Universal Grammar, frequency).

Core issues

The core questions in input processing center on the psycholinguistic conditions under which learners initially connect formal features of language with their meanings and functions, why learners make some of these connections at a given moment and not others, and what the strategies are that learners use in comprehending sentences that affect making such these connections. Underlying, of course, is the idea that connecting formal features to their meanings and functions is part of acquisition. Research on these issues led to a set of principles that guide and constrain how learners—when left to their own devices—process input data. These principles and their motivation have been described in detail in a number of other places (e.g., VanPatten, 1996, 2004, 2007, 2008) and so I will provide only a brief overview of three of them.

I begin with a definition of processing that underlies the discussion in this chapter:

Definition of processing: Processing refers to making a connection between form and meaning/ function.

What this definition underscores is that in the present model of input processing, processing is not equivalent to perception, noticing, and other terms that are often used in the SLA literature (e.g., Schmidt, 1990). As I understand these other terms, they do not require that learners connect form to meaning or function. Thus, a learner may notice a third-person -s on the end of a verb, but not connect it to any function or meaning. In short, that learner may notice but not process that form. We note here that input processing is neutral on the role of noticing; that is, it is not necessary to posit a role for noticing in the current model of input processing.

At the heart of input processing have been two key issues. The first is that input processing is tied to comprehension and the drive for learners to get meaning from the input. That is, the point of comprehension from the learner's point of view is to try to understand what someone is saying. This led to what can be called “good enough for now” processing. By “good enough for now” we mean that learners would use the easiest most efficient strategies for coping with input, especially considering that in the early stages of acquisition comprehension is likely to require a good deal of mental effort and processing resources. From this observation comes a set of principles by which learners rely heavily on lexical cues in interpreting utterances. To put it briefly, L2 learners come to the task of SLA knowing there are such things as words; that is, learners have “names for things” and can ask “how do you say X?” when asking about a word. They also know that words carry meaning. Thus, learners initially approach input processing by searchingforwordsastheeasiestwaytogetmeaning (leaving aside, for the moment, whether it is the best way to get meaning). In VanPatten (2004), I referred to this learner-centered strategy as “The primacy of content words”:

Primacy of Content Words: Learners process content words in the input before anything else.

What this principle means is not that learners know the words and are looking for them (although this may be true later in acquisition); it means that learners begin input processing in an L2 by trying to isolate words in the speech stream. While morphemes and inflections, for example, might be perceived and/or noticed, they are not processed (see definition above for processing).

So, what the principle of Primacy of Content Words suggests is that initially, learners connect lexical form to its meaning and function, but not grammatical form to its meaning or function. This idea is borne out in the acquisition literature in a number of ways. For example, in Gass's 2003 overview of interaction research, almost all research cited on the impact of interaction on learner development are vocabulary related (including the pronunciation of words that leads to misunderstanding) and Gass says that “[interaction and negative evidence] may be effective with low-level phenomena, such as pronunciation or basic meanings of lexical items” but not with morpo-syntax and certainly not with any abstractions related to syntax (Gass, 2003, p. 248). Additional evidence that learners put lexicon before morpho-syntax in a number of ways can be found in a variety of sources (e.g., Clahsen and Felser, 2006; Faerch and Kasper, 1986; Klein, 1986; Sharwood Smith, 1986; Truscott and Sharwood Smith, 2004; among others). In short, in the early stages, processing lexical items is “good enough” and gets the learners through most of the tasks they need to perform in and out of classrooms.

To be sure there are a few studies reporting learners “processing” of form before lexicon (e.g., Han and Peverly, 2007; Park, 2009), but as I have argued elsewhere (VanPatten, 2009a), such studies confuse noticing with processing (as defined here). To be sure, learners can notice just about anything, but do they connect what they notice with its meaning or function? In Park's study, for example, naïve learners of Korean report to a high degree that they kept seeing the morpheme -ta in a text they were reading. However, no learner reports that he or she knew what it meant or attached it to any function, and Park did not measure their processing. These learners simply saw it recurring in the written texts they were exposed to and noticed it.

A consequence of the Primacy of Content Words principle is another major principle, the Lexical Preference Principle. Because of my interpretation of the work of others (e.g., Carroll, 2001; Truscott and Sharwood Smith, 2004), as well as criticisms of an earlier version of this principle, (e.g., Carroll, 2004; Harrington, 2004), I have suggested that this principle ought to be stated as follows (VanPatten, 2007, 2009b):

Lexical Preference Principle: If grammatical forms express a meaning that can also be encoded lexically (e.g., that grammatical marker is redundant), then learners will not initially process those grammatical forms until they have lexical forms to which they can match them.

What this principle attempts to capture is the observable delay that learners have in processing inflectional redundancies. A good deal of parsing and processing research as well as recent advances in syntactic theory suggest that part of the “agreement problem” (i.e., concordance between items in an utterance) involves checking one item against another. L2 learners can't begin to check anything against anything unless they have the basics against which to check something. In the case of redundant grammatical markers, lexical items must be acquired before learners can acquire the grammatical markers against which they must be checked. For input processing, this means that learners wouldn't normally process grammatical markers in the input until they've processed (and acquired) corresponding lexical items first. A consequence is that the consistently redundant grammatical markers get short shrift in early stage processing because of the “good enough” strategy that works for learners: reliance on lexical items (again, defined here as content words). This idea is borne out in the literature by the fact that consistently redundant grammatical items take the longest for learners to acquire (e.g., Ellis, 1994; Gass and Selinker, 2008). To be sure, there may be other internal factors that delay the acquisition of such features, factors which are covered elsewhere in this volume (e.g., Lardiere, Chapter 7, this volume).

In addition to processing lexical items and grammatical items, learners must determine basic meanings of sentences. In Ervin-Tripp (1974), cited earlier, even when the L1 and L2 passive structures coincide word for word, L2 learners initially interpret (read “process”) passive sentences as actives. Thus, the learners tended to interpret “The lion was killed by the man” as “The lion killed the man.” In my early research on Spanish L2, I found that learners have a protracted period of misinterpreting object-verb-subject (OVS) structures as SVO. Because Spanish has much more flexible word order as well as null subjects, OVS and OV structures are possible and sometimes required. Yet learners misinterpret clitic object pronouns (and case marked fronted full noun phrases) in preverbal position incorrectly. In the example Lo vio María “Him ACC saw Mary NOM” or “Mary saw him” learners misinterpret the sentence as “He saw Mary” (see, e.g., Lee, 1987; VanPatten, 1984b). In German, case marking allows flexible word order under certain discourse and syntactic conditions. The sentence Den Jungen küsst die Frau “The boy-ACC, kisses the girl-NOM” or “The girl kisses the boy” is often misinterpreted by learners as “The boy kisses the girl” (e.g., Henry et al., 2009). Even though in natural L1 discourse such sentences might contain additional intonational cues that co-occur with these word orders, learners do not attend to them. This reliance on word order is captured by the First Noun Principle:

The First Noun Principle: Learners tend to process the first noun or pronoun they encounter in the sentence as the subject.

To be sure, some have suggested that this is not a universal processing principle but is derived from L1 transfer (e.g., Carroll, 2004; Isabelli, 2008), mainly because the vast majority of L2 learners tested have English as their L1. For example, Isabelli's 2008 study shows that Italian L1 speakers do not have the same problems interpreting Lo vio María as learners with English L1. This is because Italian is like Spanish in its use of word order and also coincides with the same lexical forms for clitic object pronouns in the singular (e.g., lo, la). Thus, they are able to make use of their L1 in processing Spanish. By extension, it is the English speakers’ L1 that gets in the way of processing Spanish because of the lack of flexible word order and OVS structures in English. This reasoning is sound but before we abandon the First Noun Principle in favor of L1 transfer, we must reconcile such findings with those of Ervin-Tripp (1974) in which L1–L2 passive equivalents did not aid L2 learners’ processing of French with English L1. We must also see what L1 Italian learners of Spanish L2, for example, do with sentences in which the L1 and L2 do not coincide.

We also know that the First Noun Principle is attenuated by lexical semantics and event probabilities, captured by the following sub-principles:

The Lexical Semantics Principle: Learners may rely on lexical semantics, where possible, instead of the First Noun Principle to interpret sentences.

The Event Probabilities Principle: Learners may rely on event probabilities, where possible, instead of the First Noun Principle to interpret sentences.

What the first sub-principle addresses is that learners bring knowledge about the semantics of verbs to the task of processing. Thus, they know that when it comes to “biting” something, the action represented by the verb requires an agent that can bite. Thus, with the two nouns “boy” and “apple” in a sentence, learners will go for “the boy” as the one who bites (irrespective of word order) because apples are incapable of performing the action. So, learners are less likely to misinterpret “The apple was bitten by the boy” compared with “The lion was killed by the man.” To be clear, this does not mean that in the “bite” sentence learners are actually processing the passive markers. They are using lexical semantics to derive meaning (see also work on the Competition Model by Bates and MacWhinney, 1984). The second sub-principle addresses what we see in acquisition in that some events are more likely than others. Returning to “bite,” if we have “rattlesnake” and “man” in the same sentence, it is more likely that the rattlesnake bites the man than the other way around, even though both are capable of biting. These two sub-principles account for the fact that sometimes learners step outside of the First Noun Principle and seemingly interpret sentences correctly not because they paid attention to grammatical cues but because they relied on non-grammatical information (see also VanPatten and Houston, 1998, for work on how informational context can attenuate processing principles).

Data and common elicitation measures

The data used for input processing research are various kinds of listening and reading measures designed to elicit how learners understand sentences and what aspects of sentences learners rely on to make determinations such as tense, aspect, and others.

Off-line measures

Off-line measures refer to those in which learners hear or read sentences and then indicate what they think the sentence means (by selecting a picture, by selecting an L1 equivalent, by choosing the most logical follow-up, by matching it to another sentence or phrase and so on). For example, learners may hear The lion was killed by the man and then asked to select from two pictures: one in which the lion is alive and the man lies dead, and another in which the man is alive and the lion lies dead. These are called off-line measures because the measurement of comprehension occurs after the learner hears the entire sentence. As another example, a learner might hear Right now the cowboy participated in the dance contest and then asked which of the following two sentences “best completes the story”: (a) He won first place. (b) He will win first place. The idea here is that the selection of a or b indicates whether the learner relied on the adverb (right now) or the verb form (participated) to assign tense to the sentence. Early research on input processing used off-line measures (e.g., Lee, 1987; VanPatten, 1984b; VanPatten and Houston, 1998). However, because off-line measures have been questioned in processing research, scholars have begun to incorporate on-line measures.

On-line measures

On-line measures refer to those in which measurements about learners’ reactions to sentences are taken at particular points during their reading or listening of sentences. Such measurements include eye-tracking, self-paced reading, and ERPs (event-related potentials). The idea behind on-line measures is that they show what learners are doing during the act of comprehension. For example, in VanPatten and Keating (2007), we tracked learners eye movements as they read sentences in which there were “tense conflicts”: sentences had inflected verbs that matched adverbials (e.g., yesterday/he played) or had inflected verbs that didn't match adverbials (e.g., right now/he played). Because we were investigating the Lexical Preference Principle which would predict that learners rely on adverbials to determine tense and not verbal inflections in the initial stages of processing, we compared native speakers of Spanish with advanced, intermediate, and beginning L2 learners. What we found was that native speakers of Spanish spend significantly more time reading verbs and not adverbs in non-match situations, suggesting a reliance on verb morphology to determine tense. Beginning and intermediate learners spent significantly more time reading adverbs rather than verbs, suggesting a reliance on adverbs to determine tense. Advanced learners patterned like native speakers. (Reading time here includes first pass on a verb/ adverb as well as subsequent additional passes, i.e., regressions.)

Applications

The basic constructs and principles of input processing have served as the basis for a pedagogical intervention known as processing instruction (sometimes referred to by others as “input processing instruction” or as simply “input processing”). The idea behind processing instruction is deceptively simple: if we know how learners approach input processing (i.e., what strategies they use, what principles underlie sentence interpretation), we might be able to construct an intervention that alters problematic processing. The result ought to be better grammatical intake for acquisition.

Although dozens of studies have now been conducted within the framework of processing instruction, the classic study often cited is VanPatten and Cadierno (1993). In that study, VanPatten and Cadierno compared two experimental groups to a control group (no instruction): processing group and a traditional group. The processing group received an input-oriented instruction in which they heard and read sentences designed to push them away from the First Noun Strategy and to interpret Spanish word order and object pronouns correctly. Thus, if they heard Lo ve la chica “The girl sees him” and incorrectly selected a picture in which the boy was looking at the girl, they were told “No, it's the other picture.” The traditional group received what was then typical foreign language instruction in the USA: a movement from mechanical to meaningful to communicative drills or practice. Thus, they practiced producing object pronouns in a variety of exercises. Two assessment tasks were given: an interpretation task that biased for the processing group and a sentence-level production task that biased for the traditional group. What VanPatten and Cadierno found was quite interesting: the processing group improved significantly on both the interpretation and the production measures, even though during the treatment they did not practice producing object pronouns. The traditional group improved on the production measure but not on the interpretation measure. There was no difference in improvement between the traditional group and the processing group on the production test. VanPatten and Cadierno concluded that somehow the processing instruction must have resulted in a change in the underlying grammar of the learners that was available for production as well as interpretation. They questioned what internal changes happened to the traditional group, although in a subsequent study (VanPatten and Wong, 2004), it seemed that the traditional group gained some kind of conscious knowledge that they applied as a test-taking strategy for the post tests—a conscious knowledge that they couldn't apply to the interpretation tests. The processing group did not show signs of such a strategy.

Since VanPatten and Cadierno's original study, a number of studies have been conducted that have examined various issues related to processing instruction. I list them here.

  • •    Processing instruction versus some other kind of intervention (e.g., traditional instruction, meaning-based output instruction, dictogloss). In sum, processing instruction is always as good as other interventions and often better (see, e.g., Benati, 2001, 2005; Cadierno, 1995; Collentine, 1998; Farley, 2001; Keating and Farley, 2008; Morgan-Short and Wood Bowden, 2006; Qin, 2008; VanPatten et al., 2009; VanPatten et al., 2009; and some of the studies in VanPatten, 2004). Some studies have contradictory findings (e.g., Allen, 2000, and Qin, 2008; VanPatten and Wong, 2004, and VanPatten et al., 2009 as well as Uludag and VanPatten, forthcoming), suggesting that replication from one study to another is difficult, especially depending on the materials and assessment tasks that researchers use.
  • •    Processing instruction and communicative assessment tasks. In this line of research, several studies have shown that the results of processing instruction appear not just on sentence-level tasks but also on discourse level tasks such as video narration and text reconstruction (e.g., Sanz and Morgan-Short, 2004; Sanz and VanPatten, 1998; VanPatten and Uludag, forthcoming).
  • •    Explicit information is not a necessary or determinate variable within processing instruction. This research has shown that providing learners with the particular kinds of activities used in processing instruction (“structured input,” e.g., Farley, 2005; Lee and VanPatten, 1995, 2003; Wong, 2005) is sufficient to cause change in knowledge and behavior. These studies include Benati (2004), Sanz and Morgan-Short (2004), VanPatten (2009c), VanPatten and Oikkennon (1996), and various studies in VanPatten (2004). Because these studies all used pre-test/post-test designs, another line of research has emerged to demonstrate that perhaps explicit information isn't necessary, but it may be beneficial during treatment. This research uses an on-line technique to track learner responses as they participate in structured input activities. The results suggest that explicit information may be utilized by learners during structured input activities, but their ability to use that information may be constrained by either structure type or the intersection of a structure with a processing problem (e.g., Culman et al., 2009; Fernández, 2008; Henry et al., 2009). Still, these on-line studies do not find that explicit information is required within processing instruction, only that it is useful.
  • •    Durative effects of processing instruction. One study has been conducted that demonstrates durable effects of processing instruction. After eight months, the effects of processing instruction were still evident; that is, final post-test scores were significantly greater than those of the pre-test, albeit there was some decline from the immediate post-test (VanPatten and Fernández, 2004). However, this is one study and replication on durative effects needs to be conducted.
  • •    Secondary effects of processing instruction. Based on a hypothesis by Lee (2004), a series of studies have shown that the effects of processing instruction on one structure transfer to other structures. For example, processing instruction on English past-tense results in increased performance on English third-person -s in the present tense. A series of these studies can be found in Benati and Lee (2008).

To be sure, there have been some criticisms of the research on processing instruction. Early criticism suggested a misunderstanding of the nature of processing instruction in that it was not clear that processing instruction aimed to alter particular processing strategies and was mistaken as just another approach to input in the classroom. Accordingly, the criticism implied that VanPatten and Cadierno (1993) were focused on “input versus output” and were somehow contextualizing grammatical structures in input (e.g., DeKeyser and Sokalski, 1996; Erlam, 2004; Salaberry, 1997; Toth, 2000; but see the response by Sanz and VanPatten, 1998 as well as VanPatten, 2002 and 2009b). We hope it is clear now that processing instruction is a particular kind of intervention that does very particular things to alter underlying processing strategies. Thus, it is quite different in both theory and intent from text enhancement, recasts, input floods, and other types of input-oriented interventions.

Subsequently, other criticisms surfaced, namely in DeKeyser et al., (2002). The major thrust of these criticisms is that (a) the model of input processing is “meaning based” rather than structurally based as L1 parsing models are, (b) processing instruction contains an inherent contradiction in that it purports to aid acquisition (an unconscious process) and yet it contains both explicit information and form-focused activities, and (c) research on processing instruction suffers from problems of internal validity. I and others have addressed these criticisms in various ways elsewhere (e.g., Carroll, 2001; Doughty, 2003; Truscott, 2004; VanPatten, 2002, 2004, 2007; VanPatten and Leeser, 2006; Wong, 2004) but in this chapter I would like to focus on the first two.

That the model of input processing is meaning based is partly true (i.e., the Lexical Preference Principle is meaning based but not the First Noun Principle). However, the question is one of how parsing in an L2 is acquired and what learners begin with. My perspective has been that learners do not begin with target-like processors and parsers but must acquire them. While there are structural aspects of parsing and processing that L2 learners adhere to (e.g., the First Noun Principle), it is also clear that the concept of “good enough” with a priority of non-syntactic processing before syntactic processing for getting data from the input is not very radical. As Clahsen and Felser (2006) have argued, the “good enough” strategy exists in both L1 and L2 processing, but L2 learners make greater use of it. What is more, the research on how learners process and interpret morphological markings in the early stages of acquisition is relatively clear: they don't. Learners rely on lexicon (and the First Noun Principle) as opposed to morphological cues to understand sentences. What learners must eventually do is move to morpho-syntactic cues and abandon a strict reliance on lexical cues.

This line of reasoning is completely in synch with other perspectives about grammatical growth (e.g., lexical categories precede functional categories); however, the focus is on how learners interact with the external data rather than how the processed data interact with, say, Universal Grammar. The reader will note that I am not juxtaposing input processing and Universal Grammar by any means. It is clear to me that input processing and Universal Grammar are both implicated in grammatical development (see, for example, the discussion in VanPatten, 1996 as well as VanPatten, 2011.)

A second criticism, an inherent contradiction in processing instruction, is that acquisition cannot make use of explicit information and form-focused tasks., The problem here centers on the definition of acquisition and what Krashen has said about the role of instruction. Krashen's claims are quite clear: (1) acquisition involves unconscious processes; and (2) explicit instruction does not affect acquisition (e.g., Krashen, 1982). Strictly speaking then, the criticism that processing instruction cannot affect acquisition as I have claimed it does would be true. But the reading of Krashen's work requires a closer inspection. First, acquisition is comprehension dependent. That is, acquisition happens in one way: learners get input, they comprehend it, and through the processing of the data during comprehension, acquisition happens (other internal/external factors ceteris paribus). This is precisely what processing instruction aims to do. It improves comprehension by making sure comprehension is correct. By making sure comprehension is correct, one impacts the data that learners make available to other internal processors and knowledge sources (e.g., Universal Grammar). In short, any instruction that fosters good comprehension ought to foster acquisition. There is no contradiction in this position.

Second, Krashen's claim that explicit instruction doesn't aid acquisition refers to two ideas: (1) explicit knowledge cannot be turned into acquisition via practice; and (2) production practice does not lead to acquisition—only the processing of input leads to acquisition. Thus, what Krashen has always argued against is form without meaning, practice as production, the need for explicit information, and what generally characterized instruction at the time of his writings in the late 1970s and throughout the 1980s. Processing instruction is a direct reaction to that same kind of instruction, and it does not purport to do any of the things that Krashen railed against in his condemnations of language teaching. For this reason, early on I characterized processing instruction as an intervention for the “acquisition rich” classroom (VanPatten, 1993). What is more, as we have shown in at least one line of research within processing instruction, explicit information is not even needed in processing instruction (see above); appropriately crafted structured input alone is sufficient to cause the changes in learner knowledge and performance we see in our research. (For additional arguments on the implications for instruction from an acquisition/processing perspective, see Truscott and Sharwood Smith, 2004).

Future directions

One area in which input processing will change or evolve is in part determined by what I call the difference between processing as acquisition and processing as a by-product of acquisition (VanPatten and Jegerski, 2011). Processing as acquisition refers to how learners comprehend and process data in the input to construct linguistic systems. Processing as a byproduct refers to how learners make use of that linguistic system to comprehend utterances. I will illustrate with a simple example from null subjects in Spanish. Unlike English, Spanish is a null subject language, meaning that in simple declarative sentences overt subject pronouns are not necessary (e.g., Habla/Él habla “He speaks” are both grammatical) and in certain kinds of sentences overt subjects are disallowed (e.g., time expressions such as Es la una/*Ello es la una “It's one o'clock,” weather expressions such as Está lloviendo/Ello está lloviendo “It's raining” and others). Processing for acquisition refers to how learners get data from the input that then informs the internal system that Spanish is a null subject language. This, of course, is relatively straightforward in that learners encounter early on subjectless sentences and the grammar is forced to posit Spanish as a null subject language. In short, the absence of an overt subject motivates projecting null subjects.

But once the grammar is established, how do learners subsequently interpret null and explicit pronouns they encounter in the input? This is processing as a by-product of acquisition in that the learner is no longer processing to create a system but is learning to equate new pragmatic meanings with the existing syntax. How does the learner interpret, for example, the embedded clauses in the two sentences Juan vio a Roberto después que, regresó de Europa (“John saw Robert after [null subject] returned from Europe”) and Juan vio a Roberto después que él regresó de Europa (“John saw Robert after [él = ‘he’, explicit subject] returned from Europe”). Do learners distinguish between antecedents (e.g., John vs. Robert) when the pronoun in the second clause is either null or overt? Spanish speakers seem to but it is not clear that learners do (e.g., Sorace and Filiaci, 2006); but again, this is not related to the recreation of a linguistic system that permits null subjects. It is related to what learners do with a null and explicit pronoun system once they get it. (For a summary of work on null subjects and Spanish, see Montrul, 2004.)

The point of this distinction is that input processing cannot address the issue of what happens after grammatical features are acquired. It is not the point of input processing to address the resolution of, for example, ambiguity during comprehension. The point of input processing is to address how learners create a grammar. To be sure, the distinction between processing for acquisition and processing as a by-product of acquisition may not be so categorical for all areas of language (again, see the discussion in VanPatten and Jegerski) but I believe it is a useful distinction to keep in mind as we evaluate various models of processing and parsing applied to the L2 context.

This leads me to one future direction of input processing. In addition to the current principles it contains, input processing may have to branch out and include features of parsing. In the case of null subjects, in order for learners to grasp that Spanish is a null subject language, they must process the absence of an overt subject in the input. This can only be done by recurring to principles of parsing such as those in Pritchett (1992). According to Pritchett, during processing the parser attempts to meet the requirements of the theta criterion, that is, find and tag all required thematic elements of the verb (e.g., agent, theme, goal) in the sentence. If there is no overt subject in a sentence that corresponds to an underlying thematic element of the verb, then the parser must posit a null subject in order for the parse of the sentence to be accepted. Thus, learners at some point in SLA have to make use of such principles for a complete grammar to develop. In the case of null subjects, there is no foreseeable way that the learner would posit null subjects unless there was something in the parser that demanded that all thematic elements of the verb be represented somewhere in the sentence. I see, then, that input processing will expand to include not only those principles that currently comprise it, but others as well. These other principles may be some of those currently found in parsing models. I do not believe, however, that input processing will be replaced by parsing models. Again, such models do not address the initial drive to get lexicon before grammatical inflection, for example. Instead, I see input processing and parsing merging in some way to better account for how a grammar is acquired.

I also see an area of research that is still open to question; that is, the role of L1 in input processing. As currently formulated, the principles of input processing are universal in nature and apply to all learners of all languages in all contexts. Carroll (2001), on the other hand, has been a strong proponent of L1 processing underlying all initial L2 processing. The question, then is this: Do L2 learners transfer L1 processing routines and strategies to SLA? Or, is L2 input processing guided by universal principles? Alternatively, is there some mix of the two, such as, do learners start with universal principles but L1 processing routines “kick in” when triggered by certain data in the input? Some research on sentence processing more generally has suggested strong L1 influence (e.g., Frenck-Mestre, 2005; Serratrice, 2007; Sorace and Filiaci, 2006) while other research has not (e.g., Marinis et al., 2005; Papadopoulou and Clahsen, 2003). However, these studies examine processing as by-product (how learners make use of grammar while processing) and thus do not address the initial stages of processing as acquisition (how learners get the grammar). In short, I see the role of L1 in processing for acquisition open for discussion and continued research.

Conclusion

Since the early 1990s, input processing has enjoyed a rigorous research agenda, in part due to the immediate instructional applications that arose (e.g, VanPatten and Cadierno, 1993). Like other aspects of acquisition, its relative merits have been questioned and debated, and it is clear that input processing will evolve. In some ways, it already has. For example, Truscott and Sharwood Smith's (2004) acquisition as processing subsumes/assumes the current model of input processing in that their perspective provides a much larger view of acquisition that includes processing. Yet, at the same time, their theory does not detail particular strategies that may or may not impede the processing of input data, as their main concern is overarching theoretical issues concerning how grammars change over time. Because of the strong line of application research that has emerged from input processing—namely processing instruction—it is likely that some version of the current model of input processing will remain in the mainstream for some time, perhaps side by side with other approaches to acquisition as processing and processing for acquisition.

References

Allen, L. (2000). Form-meaning connections and the French causative: An experiment in processing instruction. Studies in Second Language Acquisition, 22, 69–84.

Bates, E. and MacWhinney, B. (1984). Functionalism and the Competition Model. In B. MacWhinney and E. Bates (Eds.), The cross-linguistic study of sentence processing (pp. 3–73). Cambridge: Cambridge University Press.

Benati, A. (2001). A comparative study of the effects of processing instruction and output-based instruction on the acquisition of the Italian future tense. Language Teaching Research, 5, 95–127.

Benati, A. (2004). The effects of structured input activities and explicit information on the acquisition of the Italian future tense. In B. VanPatten (Ed.), Processing instruction: Theory, research, and commentary (pp. 207–225). Mahwah, NJ: Lawrence Erlbaum and Associates.

Benati, A. (2005). The effects of processing instruction, traditional instruction, and meaning-output instruction on the acquisition of the English past simple tense. Language Teaching Research, 9, 67–93.

Benati, A. and Lee, J. F. (2008). Grammar acquisition and processing instruction. Bristol, UK: Multingual Matters.

Bever, T. G. (1970). The cognitive basis for linguistic structures. In J. R. Hayes (Ed.), Cognition and the development of language (pp. 279–362). New York: John Wiley and Sons.

Cadierno, T. (1995). Formal instruction from a processing perspective: An investigation into the Spanish past tense. The Modern Language Journal, 79, 179–93.

Carroll, S. (2001). Input and evidence: The raw material of second language acquisition. Amsterdam: John Benjamins.

Carroll, S. (2004). Commentary: Some general and specific comments on input processing and processing instruction. In B. VanPatten (Ed.), Processing instruction: Theory, research, and commentary (pp. 293–309). Mahwah, NJ: Lawrence Erlbaum Associates.

Clahsen, H. and Felser C. (2006). Grammatical processing in language learners. Applied Psycholinguistics, 27, 3–42.

Collentine, J. (1998). Processing instruction and the subjunctive. Hispania, 81, 576–587.

Culman, H., Henry, N., and VanPatten, B. (2009). The role of explicit information in processing instruction: An on-line study with German accusative case inflections. Die Unterrichtspraxis.

DeKeyser, R., Salaberry, M. R., Robinson, P., and Harrington, M. (2002). What gets processing in processing instruction? A response to Bill VanPatten's “Update”. Language Learning, 52, 805–823.

DeKeyser, R. and Sokalski K. (1996). The differential role of comprehension and production practice. Language Learning, 46, 613–642.

Doughty, C. J. (2003). Instructed SLA: Constraints, compensation, and enhancement. In C. J. Doughty and M. H. Long (Eds.), The handbook of second language acquisition (pp. 256–310). Oxford: Blackwell.

Ellis, R. (1994). The study of second language acquisition. Oxford: Oxford University Press.

Erlam, R. (2004). Evaluating the effectiveness of structured input and output-based instruction in foreign language learning. Studies in Second Language Acquisition, 25, 559–582.

Ervin-Tripp, S. M. (1974). Is second language acquisition really like the first?. TESOL Quarterly, 8, 111–127.

Faerch, C. and Kasper G. (1986). The role of comprehension in second language learning. Applied Linguistics, 7, 257–274.

Farley, A. P. (2001). Authentic processing instruction and the Spanish subjunctive. Hispania, 84, 289–299.

Farley, A. P. (2005). Structured input. New York: McGraw-Hill.

Fernández, C. (2008). Reexamining the role of explicit information in processing instruction. Studies in Second Language Acquisition, 30, 277–305.

Frenck-Mestre, C. (2005). Eye-movement recording as a tool for studying syntactic processing in a second language: A review of methodologies and experimental findings. Second Language Research, 21, 175–198.

Gass, S. M. (2003). Input and interaction. In C. Doughty and M. H. Long (Eds.), The handbook of second language acquisition (pp. 224–255). Oxford: Blackwell.

Gass, S. M. and Madden, C. G. (Eds.) (1985). Input in second language acquisition. Rowley, MA: Newbury House.

Gass, S. M. and Selinker, L. (2008). Second language acquisition: An introductory course. (Third Edition). New York: Routledge.

Han, Z. and Peverly, S. T. (2007). Input processing: A study of ab initio learners with multilingual backgrounds. International Journal of Multilingualism, 4, 17–37.

Harrington, M. (2004). Commentary: Input processing as a theory of processing input. In B. VanPatten (Ed.), Processing instruction: Theory, research, and commentary (pp. 79–92). Mahwah, NJ: Lawrence Erlbaum Associates.

Hatch, E. (1983). Simplified input. In R. W. Andersen (Ed.), Pidginization and creolization as language acquisition (pp. 64–88). Rowley, MA: Newbury House.

Henry, N., Culman, H., and VanPatten, B. (2009). More on the effects of explicit information in instructed SLA: A partial replication of and a response to fernández (2008). Studies in Second Language Acquisition, 31, 559–575.

Isabelli, C. (2008). First Noun Principle or L1 transfer principle in SLA? Hispania, 2, 463–476.

Keating, G. D. and Farley, A. P. (2008). Processing instruction, meaning-based output instruction, and meaning-based drills: Impacts on classroom L2 acquisition of Spanish object pronouns. Hispania, 91, 631–650.

Klein, W. (1986). Second language acquisition. Cambridge: Cambridge University Press.

Krashen, S. D. (1982). Principles and practice in second language acquisition. Oxford: Pergamon Press.

Lee, J. F. (1987). Morphological factors influencing pronominal reference assignment by learners of Spanish. In T. A. Morgan, J. F. Lee, and B. VanPatten (Eds.), Language and language use: Studies in Spanish (pp. 221–232). Lanham, MD: University Press of America.

Lee, J. F. (2004). On the generalizability, limits, and potential future directions of processing instruction research. In B. VanPatten (Ed.), Processing instruction: Theory, research, and commentary (pp. 311–323). Mahwah, NJ: Lawrence Erlbaum and Associates.

Lee, J. F. and VanPatten, B. (1995). Making communicative language teaching happen ( First Edition). New York: McGraw-Hill.

Lee, J. F. and VanPatten, B. (2003). Making communicative language teaching happen (Second Edition). New York: McGraw-Hill.

Long, M. (1983). Native speaker/non-native speaker conversations and the negotiation of comprehensible input. Applied Linguistics, 4, 126–141.

Marinis, T., Roberts, L., Felser, C., and Clahsen, H. (2005). Gaps in second language processing. Studies in Second Language Acquisition, 27, 53–78.

Montrul, S. (2004). The acquisition of Spanish: Morphosyntactc development in monolingual and bilingual L1 acquisition and adult L2 acquisition. Amsterdam: John Benjamins.

Morgan-Short, K. and Wood Bowden, H. (2006). Processing instruction and meaningful output-based instruction: Effects on second language development. Studies in Second Language Acquisition, 28, 31–65.

Papadopoulou, D. and Clahsen, H. (2003). Parsing strategies in L1 and L2 sentence processing: A study of relative clause attachment in Greek. Studies in Second Language Acquisition, 25, 501–528.

Park, E. (2009). Learners’ self-generated noticing of L2 written input: noticed features and processing strategies. Paper delivered at the annual meeting of the Georgetown University Roundtable on Linguistics.

Pritchett, B. L. (1992). Grammatical competence and parsing performance. Chicago: University of Chicago Press.

Qin, J. (2008). The effect of processing instruction and dictogloss tasks on acquisition of the English passive voice. Language Teaching Research, 12, 61–82.

Salaberry, M. R. (1997). The role of input and output practice in second language acquisition. The Canadian Modern Language Review, 53, 422–451.

Sanz, C. and Morgan-Short, K. (2004). Positive evidence vs. explicit rule presentation and explicit negative feedback: A computer assisted study. Language Learning, 54, 35–78.

Sanz, C. and VanPatten, B. (1998). On input processing, processing instruction, and the nature of replication tasks: A response to Salaberry. The Canadian Modern Language Review, 54, 263–273.

Schmidt, R. W. (1990). The role of consciousness in second language learning. Applied Linguistics, 11, 129–158.

Serratrice, L. (2007). Cross-linguistic influence in the interpretation of anaphoric and cataphoric pronouns in English-Italian bilingual children. Bilingualism: Language and Cognition, 10(3), 225–238.

Sharwood Smith, M. (1986). Comprehension versus acquisition: Two ways of processing input. Applied Linguistics, 7, 239–274.

Slobin, D. I. (1966). Grammatical transformations and sentence interpretation in childhood and adulthood. Journal of Verbal Learning and Verbal Behavior, 5, 219–227.

Sorace, A. and Filiaci, F. (2006). Anaphora resolution in near-native speakers of Italian. Second Language Research, 22(3), 339–368.

Toth, P. D. (2000). The interaction of instruction and learner-internal factors in the acquisition of L2 morphosyntax. Studies in Second Language Acquisition, 22, 169–208.

Truscott, J. (2004). The effectiveness of grammar instruction: Analysis of a meta-analysis. English Teaching and Learning, 28, 17–29.

Truscott, J. and Sharwood Smith, M. (2004). Acquisition by processing: A modular perspective on language development. Bilingualism: Language and Cognition, 7, 1–20.

Uludag, O. and VanPatten, B. (forthcoming). The comparative effects of processing instruction and dictogloss on the acquisition of the English passive by speakers of Turkish. International Review of Applied Linguistics.

VanPatten, B. (1984a). Processing strategies and morpheme acquisition. In F. R. Eckman, L. H. Bell, and D. Nelson (Eds.), Universals of second language acquisition (pp. 88–98). Rowley, MA: Newbury House.

VanPatten, B. (1984b). Learners’ comprehension of clitic pronouns: More evidence for a word order strategy. Hispanic Linguistics, 1, 57–67.

VanPatten, B. (1993). Grammar instruction for the acquisition-rich classroom. Foreign Language Annals, 26, 433–450.

VanPatten, B. (1996). Input processing and grammar instruction. Norwood, NJ: Ablex.

VanPatten, B. (2002). Processing the content of input processing and processing instruction: A response to DeKeyser, Salaberry, Robinson, and Harrington. Language Learning, 52, 825–831.

VanPatten, B. (Ed.) (2004). Processing instruction: Theory, research, and commentary. Mahwah, NJ: Lawrence Erlbaum Associates.

VanPatten, B. (2007). Input processing in adult second language acquisition. In B. VanPatten and J. Williams (Eds.), Theories in second language acquisition (pp. 115–135). Mahwah, NJ: Lawrence Erlbaum Associates.

VanPatten, B. (2008). Processing matters. In T. Piske and M. Young-Scholten (Eds.), Input Matters (pp. 47–61). Clevedon, UK: Multilingual Matters.

VanPatten, B. (2009a.) Stubborn syntax. Plenary paper delivered at the annual meeting of the Georgetown University Roundtable on Linguistics.

VanPatten, B. (2009b). Processing matters in input enhancement. In. T. Piske and M. Young-Scholten (Eds.), Input matters (pp. 47–61). Bristol, UK: Multilingual Matters.

VanPatten, B. (2009c). Formal intervention and the development of proficiency: The role of explicit information. In A. Benati (Ed.), Issues in attaining second language proficiency (pp. 169–188). London: Continuum.

VanPatten, B. (2011). Stubborn syntax: How it resists explicit teaching and learning. In C. Sanz and R. Leow (Eds.), Proceedings of the 2009 Georgetown University Roundtable on Linguistics. Washington, D.C.: Georgetown University Press.

VanPatten, B. and Cadierno, T. (1993). Explicit instruction and input processing. Studies in Second Language Acquisition, 15, 225–243.

VanPatten, B. Farmer, J. L., and Clardy, C. L. (2009). Processing instruction and meaning-based output instruction: A response to Keating and Farley (2008). Hispania, 92, 116–126.

VanPatten, B. and Fernández, C. (2004). The long-term effects of processing instruction. In B. VanPatten (Ed.), Processing instruction: Theory, research, and commentary (pp. 273–289). Mahwah, NJ: Lawrence Erlbaum Associates.

VanPatten, B. and Houston, T. (1998). Contextual effects in processing L2 input sentences. Spanish Applied Linguistics, 2, 53–70.

VanPatten, B., Inclezan, D., Salazar, H., and Farley, A. P. (2009). Processing instruction and dictogloss: A study on word order and object pronouns in Spanish. Foreign Language Annals, 42, 558–576.

VanPatten, B. and Jegerski, J. (2011). L2 processing and parsing: The issues. In B. VanPatten and J. Jegerski (Eds.), Second language processing and parsing (pp. 3–23). Amsterdam: John Benjamins.

VanPatten, B. and Keating, G. D. (2007). The hidden problem of second language acquisition: Learners’ processing of morpho-syntax during comprehension. Paper delivered at the annual meeting of the American Association of Teachers of Spanish and Portuguese, San Diego.

VanPatten B. and Leeser, M. J. (2006). Theoretical and research considerations underlying classroom practice: The fundamental role of input. In R. Salaberry and B. Lafford (Eds.), The art of teaching Spanish: Second language acquisition from research to praxis (pp. 55–77). Georgetown: Georgetown University Press.

VanPatten, B. and Oikkenon, S. (1996). Explanation versus structured input in processing instruction. Studies in Second Language Acquisition, 18, 495–510.

VanPatten, B. and Sanz, C. (1995). From input to output: Processing instruction and communicative tasks. In F. R. Eckman, D. Highland, P. W. Lee, J. Milcham, and R. Ruthkowski Weber (Eds.), Second language acquisition theory and pedagogy (pp. 169–85). Mahwah, NJ: Lawrence Erlbaum Associates.

VanPatten, B. and Uludag, O. (2011). Transfer of training and processing instruction: From input to output. System, 39, 44–53.

VanPatten, B. and Wong, W. (2004). Processing instruction and the French causative: Another replication. In B. VanPatten (Ed.), Processing instruction: Theory, research, and commentary (pp. 97–118). Mahwah, NJ: Lawrence Erlbaum Associates.

Wong, W. (2004). The nature of processing instruction. In B. VanPatten (Ed.), Processing instruction: Theory, research, and commentary (pp. 33–63). Mahwah, NJ: Lawrence Erlbaum Associates.

Wong, W. (2005). Input enhancement: From theory and research to classroom practice. New York: McGraw-Hill.