Error Correction in Language Teaching

  • This essay is a short review of one specific aspect of Brown and Larson-Hall’s 2012 introductory book Second Language Acquisition Myths. In particular, this review addresses the myth of error correction, which Brown and Larson-Hall phase as “Language Learners Always Benefit From Error Correction”.

Introduction

The question of error correction is certainly perplexing to many of us language teachers. Not only which form of correction, but even how to correct children as opposed to adults. As a teacher cited in Brown and Larson-Hall’s (2012) Second language acquisition myths  says, “[c]hildren make adorable mistakes” (p. 105; italics mine). Adults generally do not make adorable mistakes. The intersection between age and correction-type then, is in the center of this so-called myth. In my personal experience however, it seems the authors miss a few important factors in their deconstruction of this myth.

Beyond declaring one type of error correction better than another is the idea of dynamic correction (or assessment) (Lantolf & Poehner, 2011). Additionally, this myth is ripe for native-speaker bias in which native-speaking teachers correct what they perceive as errors, but what may actually be a more creative act of self-expression. Finally, while perhaps beyond the scope of their book, no mention is made of the philosophical concerns related to evaluation, both negative and positive. The categorizing and approving of specific types of error correction, which is the hallmark of the science cited in the chapter, seems to miss what I have experienced in my own classrooms as a more messy relationship between learners and educators and the process of evaluation.

Error Correction

Prologue: Native-speaker bias

Native-speaker knows best

 

To begin this discussion, I want to first address a problem present in the book, which goes largely unaddressed. In the chapter on error correction, the teacher “in the real world” (p. 105) points out an English construction they view as a grammar mistake of helping verbs and their objects. The teacher identifies the construction, “Let’s English!” as grammatically incorrect, because let requires a main verb (e.g. “Let’s do/eat/finish it!”). The teacher expresses the frustration of many when she says, “I have to admit that sometimes I just give up and say, OK let’s English! Corrections seem to do no good in some cases” (Brown & Larson-Hall, 2012, p. 106).

Verbing nouns is not uncommon in so-called native English. In my own classroom, we often play the card game Coup (Indieboardsandcards, 2016), where each player has two cards with names of certain kinds of people on them, like Duke, assassin, captain and so on. Each of these cards can perform certain actions in the game, like tax or exchange. When performing your action, it seems reasonable that you would use the verb to declare your action (e.g. “I will tax and take two coins!”). However, after a few games I, their teacher, began saying, “I duke your foreign aid”. instead of “[As a duke], I block your foreign aid”. The students soon caught on and began verbing the names of all the cards.

As a native speaker, I seemingly have the authority to be creative with my so-called errors. However, had my students begun doing the same thing without my help, would I have corrected them? I hope not, or at least I hope I would figure out a better way to understand why they did it. If it was a genuine mistake, I suppose I would want to help fix it. But if it was self-expression and creativity, well then why do I have the right to self-express, but not my language-learning students? The grayness of the two is nearly impossible to tease apart.

Recasts vs prompts

Brown and Larson-Hall (2012) end their chapter on error correction by advising teachers to use explicit prompts which both draw the attention of the learner to their error and provide an opportunity for them to correct it. Especially for low-level learners (and younger learners), explicit correction makes it clear what the teacher is expecting. This conclusion follows from the research provided, where two different types of error correction (recasts and prompts) are studied in experimentally sound conditions and generally prompts produce the kind of language improvement teachers look for.

Recasts, the authors note, are the preferred type of error correction by many teachers. They seem less intrusive to the language production of the learner and less directly evaluative. However, students, especially young learners, don’t seem to pick up on grammar corrections if they are not obviously pointed out.

This follows from the experience of everyone who has learned a language. The only time recasts happen naturally in the world is when interlocutors A) physically can not hear the speaker or B) the meaning of the word or sentence was unclear. In other words, recasts are communicative repair tools that people use to understand the meaning of each other’s speech. Additionally, it is generally found to be very rude and pompous to correct another person’s grammar in this way.

Explicit prompts, on the other hand, interrupt the learner’s communicative act, in order to step-down and address the grammar. I find that this interruption, especially in the beginning, may be surprising or annoying to students, but the research suggests that bringing learners’ attention directly to the problem helps them notice it. As long as the student is paying any attention to their teacher, their mental resources have to switch from communication to logical thinking about language.

This categorizing and evaluating error correction types seems to show that, of course, explicit prompts are better. However, we need to think more clearly about what our categories do. In many enterprises, category formation focuses specific content as exemplary and tends to ignore content on the boundary. When we think of error correction strategies then, it seems that explicit prompts are the best fit for the category.

Dynamic assessment

There is another way to think about error correction however. Instead of categories, we can think about error correction continuums, where the type of error correction selected by the teacher should meet the need of the learner. Instead of simply using “more” of one type of correction over the other. I believe that dynamic assessment (DA) (Lantolf & Poehner, 2011) provides a better method for error correction than simply preferring explicit prompts over recasts.

DA can allow teachers to see more clearly what level their learners are at, beyond just “can do X, can’t do Y”. It asks teachers to address a learner at the stage they are at. If a learner makes a mistake, but has almost nearly mastered the content, it may only take a questioning look from the teacher to make the student realize their mistakes. Other students may not even know that they don’t know they made a mistake. The point is, the teacher needs to listen to the student, identify where they fail to address their problem, and give correction at a scaled level. This changes the suggestion we would give teachers. It’s not simply be more explicit, it’s: meet your learner where they are.

Evaluation

Negative and positive evaluation

In my view however, this discussion of good, better, best jumps ahead to quickly. No mention in the chapter is directed at the concept of correction fundamentally. At a basic level, correction is a type of evaluation that goes beyond identifying good or bad production to providing opportunity to correct it. However, acknowledging the root of evaluation is important because it raises important philosophical implications for our language teaching and learning.

Earl Stevick in his language teaching book A Way and Ways (1980) notes that while negative evaluation is at times seen as controversial, positive evaluation is more often assumed to be good no matter what. This is seen as a mistake and it is argued that positive evaluation can be as harmful as negative. Stevick (1980) describes the use of evaluation as establishing an “evaluative climate” (p. 23) in the classroom. This climate is a tension between what is called the performing self and the evaluative self that resides in all of us. This is linguistically evident in phrases like “I’m my own toughest critic”. Often that is the truth. However, the same tension that we all at times inwardly feel is often unknowingly established in the language classroom through any type of evaluation.

Stevick (1980) illustrates this idea by an experience he had with a german friend who once told him, “Oh, I like talking with you. You use such correct grammar” (p. 23). This immediately made him self-conscious of his grammar, whereas he wasn’t before. I also felt this same tension while talking to people in Brazil. Often after “hello”, people would comment “Wow! you speak such good Portuguese!”. This experience always led me to feel pressure, because I knew that after a few minutes of talking, they would realize I didn’t actually speak such good Portuguese. I just happened to say “hello” very naturally according to them. This experience happens much more often in Korea, where any knowledge of Korean by a western-foreigner is met with much praise and adoration.

Natural evaluation in language

The issue with positive evaluation in the classroom, I think, lies in the way we naturally express positivity towards the language use of other people when they speak. In a phrase, we don’t. We signal satisfaction with the words of other people by attending more and responding to their language use. When someone says something interesting, we say, “That’s (the content)  interesting!” or “Well said (the content)!”. It would be seen as patronizing to say, “Very good on your grammar”.

Brown and Larson-Hall (2012) acknowledge this idea when they talk about the teacher who reacts communicatively to the content of their learner’s mistakes of “have” and “be” and this is how I try to signal error correction in my own classroom. I try to show interest in the content, while also signaling that the error is in the language itself, not the semantics. It is a difficult thing to do.

Conclusion

The word myth is something of a sledgehammer. It lacks nuance. It certainly is false that learners always learn from error correction; however, it is a very strangely worded myth. Brown and Larson-Hall (2012) handle the scientific debate between different types of error correction in a way that is probably helpful to many language teachers, but I wonder if they don’t in some ways only change the problem quantitatively and miss a greater opportunity to really address the issue with correcting language learner errors.

The scientific categorization of error correction types divorces the ecological practice of teachers and the needs of students. It says “this is better quantitatively” without perhaps addressing the qualitative issues and fears of teachers and students in evaluation.


References

Brown, S., & Larson-Hall, J. (2012). Second language acquisition myths. Ann Arbor, MI: University of Michigan Press. 

indieboardsandcards. (2016). Indieboardsandcards.com. Retrieved 5 March 2016.

Lantolf, J., & Poehner, M. (2010). Dynamic assessment in the classroom: Vygotskian praxis for second language development. Language Teaching Research, 15(1), 11-33.

Stevick, E. W. (1980). Teaching languages: A way and ways. Rowley: Newbury House Publishers.

Game-Design Enhanced Language Teaching

Let me start by saying this: “Game-design” language teaching is not the gamification of teaching. It is not “playing games” in a class in order to boost “fun”, though that is certainly a benefit. Game-design language teaching is something much bolder. (Check here for a brief definition of the differences). What follows here are a few principles of game-design and how they can be leveraged to teach languages in ways that address some of the hardest problems in language teaching. Namely, authentic communities of practice, learner-driven tasks and Dynamic, just-in-time feedback. This article will lay the foundation for future, praxis-oriented, content such as specific games for language teaching and specific language-learning plans.

 

Introduction

zet8tzb

Although, for me, Age empires would have been blue

Games in language teaching are, without question, seen as an integral and essential part of most teachers’ lesson plans. The intensity or involvement in any given game varies from teacher to teacher, class to class. From language puzzles like crosswords and hangman to active Dungeons and Dragons-esque RPGs. Researchers, too, have for decades recommended games in the language classroom for various reasons including the development of positive attitudes towards learning, providing clear goals and engagement (Palmer & Rodgers, 1983). Games, however, often only serve one specific goal for teachers: student engagement. Games are fun. The rise of gamification is related to this problem and it is an attempt to solve the same kinds of problems.

For this reason, there often is not much thought or discussion given to why we should play games, or how the games are beneficial language learning tools. Anecdotally, many teachers like myself know that games engage students and help scaffold their language production. But still, how? If it is true that games fundamentally help language learning, what principles of games and game design lead to these kinds of outcomes? These are the kinds of questions that researchers like Julie Skyes and James Paul Gee have begun to address. What follows here is a brief introduction and foundation to ground theoretically what will hopefully become a repository of useful information regarding games for second language learning and teaching (L2TL).

While many researchers have examined the language learning effects from games in themselves, as teachers many of us are interested in the ways we can effectively coordinate and implement games, particularly digital games relevant to our “digital native” (Prensky, 2003) students, in and around our classrooms. While games in general may be found to be effective language learning tools, how game-design, and in particular digital game-design can be leverage to improve pedagogy and classroom outcomes.

Some Assumptions

Language socialization 

Before going forward, I want to briefly mention some of the underlying theory and philosophy that leads many researchers and teachers like myself to believe that there is real value in using games to teach language. A game-design approach to L2TL takes a principled approach based on socio-cultural theories of language and teaching, in particular, language socialization (LS) (Ochs & Schieffelin, 1995) and tasked-based language teaching (TBLT) (Willis, 1996). An LS perspective views language and language learning as means-end oriented. That is, children come to learn and use language as a tool, or means, to achieve some goal, or end. These means and ends from an LS perspective are enforced by culture.

Ochs and Schieffelin’s perhaps most well-known example relates to how children are addressed differently in different cultures. They note that middle-class American caregivers are very likely to spend a lot of time directing their speech to their pre-linguistic infants. The speech is not random however. Nor is it monologue. It is conversational. Caregivers in these communities greet, ask questions and correct behavior all the while knowing the child can neither understand nor respond to them. Additionally, these caregivers go above and beyond normal conversational behavior to maintain the attention of the infant by exaggerating their tone of voice and body gesture.

Ochs and Schieffelin then go on to describe a different cultural practice by some Mayan (and many other) communities. Caregivers in these communities do not interact directly with pre-verbal infants. Adults in these communities do not view infants as appropriate conversational partners and so do not engage them directly with language until they already know how to talk.

The brilliance of the theory of Language Socialization then, is its ability to account for both of these cultural practices. Importantly, there is no single rule that governs the language development in these cultures. It is not the case that middle-class Americans learn language by direct conversation with their caregivers, nor is it the case that Mayan babies don’t learn language. Instead, LS says that different communities organize around the central principles of acceptable participation and evolutionary-driven needs (e.g. food, attention, touch and so on) that infants need to obtain. “novices”, as the children in each case are called, operate on the periphery of language communities. They are “peripheral members”. Core members of the community are those individuals and groups who have power and influence. These are generally adults, but can be other non-center groups who are nontheless closer to the center than the infant (such as older children). For us language teachers, it is easy to recognize that our L2 learners are peripheral members of their L2 language communities too.

So, how do the infant-directed approaches to language fit into this system of center and peripheral membership? For the Mayan children, they are required to follow their mother where ever she goes. The communal aspect of the every day life, on average, means that the child will be exposed to a great deal of language and conversation between their caregiver and other adults in the community. In the terms of Ochs and Schieffelin, the children become overhearers. They are legitimate peripheral members of the community. They are welcome to be present. For the American middle-class child however, particularly in traditional nuclear families, the child may spend days in the company of primarily one adult. The child in this situation then, is not overhearing much language at all. And hence, the effort of the caregiver to directly engage the pre-language infant.

For us, the L2TL educator, their are many other important insights that the LS perspective brings, but I just want to highlight the importance of this initial insight. Language learning is always the result of socializing into a community. And to do so, it is necessary to move from the periphery of that community towards the center. For our L2TL students then, helping them gain a legitimate and authentic position as a peripheral member of a community is a crucial step.

It is often assumed that the language classroom itself is that legitimate and authentic community. However, the goal of the language classroom is not to create a community of language classroom L2 users. But instead, to transfer the skills developed and honed in the classroom into a real community of speakers. Creating a positive and welcoming language classroom community is a real and important goal, but it is not itself the community we are aiming for. At least because the language classroom is necessary transitional and temporary, necessarily unstable.

  • So, we see a few important insights, as well as important challenges for L2TL. Membership in a language community is vital, but how to do so in a mostly inauthentic classroom setting?

Task-based Learning and Teaching (TBLT)

image002

The general structure of a TBLT lesson

Traditionally, TBLT was motivated by the failures of methods such as audio-lingual  and Communicative Language Teaching (CLT). Importantly, how each of those methods failed to incorporate vital aspects of the other. In the case of the grammar-translation and audio-lingual, it was the complete lack of social knowledge needed to actually use the grammar rules or pronunciations learned through those methods. And in the case of CLT, the fact that students seemed to make great gains in comprehension of content language, but failed to make similar gains in functional aspects of grammar. TBLT attempts to fix these problems and also motivate students, but marrying real-life language use with problems that needed to be solved through language.

TBLT, in relation to LS, can be seen as a pedagogical means-end approach to L2TL. Sykes and Reinhardt (2013) situate the theoretical foundation of TBLT in communicative language teaching related to functional approaches to language. From this view, it was important that L2TL be applicable to real-world situations and not just the learning of language rules and forms. This end, communication in real-world situations, is achieved through pedagogically valid tasks that act as means to those ends.

A classic example is a hotel checklist task. Originally designed for immigrant ESL learners who had been employed in hotels as maids or other staff. The task, following the diagram above, gives a pair of students the task of cleaning a hotel room. Each learner has a different set of tasks that they have either “done” or “not done”. Most of the tasks, between the two learners, have been done. The goal of the learners is to determine what still needs to be done to successfully prepare the hotel room.

This task has the benefit of being immediately applicable to the learners, it involves words and language structures that will be immediately important and useful to them (words related to working in a hotel). Additionally, the learners must negotiate together to recognize and notice BOTH the content (what still needs to be done) and also the LANGUAGE (by carefully reading through their lists). There are also several problems with tasks like this, and if you’ve been teaching, you may be able to recognize any number of them. What follows is a short list (not comprehensive) of some of the problems that TBLT sometimes face, and how game-design can help us overcome them.

A critique of TBLT from a Game-Design Perspective

For L2TL at the moment, TBLT is the target most teachers are trying to hit. And there are many different approaches, pedagogues and lesson plans that utilize TBLT. However, some have recognized several short-comings in TBLT. In particular, researchers interested in game-design have discovered a short list of items that could be improved, and ways to do it.

Sykes and Reinhardt’s perspective of game-informed language socialization and task-based L2TL are categorized in five problematic areas:

  1. goals and tasks – lack of student agency
  2. interaction – promotes good ideational (semantic) interaction, but can fail to promote adquate interpersonal interaction
  3. feedback – Delayed, hard to give the correct feedback in the moment it will help
  4. context and narrative – Tasks can often be divorced from broader societal narratives and context that drives language learning (from an LS perspective).
  5. motivation – Tasks may motivate for awhile, but often do not lead to sustained, motivated attention over multiple class sessions..

A game-design TBLT perspective views each of the five areas of criteria differently than traditional views might. For this reason, we will briefly introduce them in this section along with a description of what good games do with the five TBLT areas mentioned above.

Sykes and Reinhardt describe goals and tasks in game-designed TBLT as centering around two main ideas: having learner driven tasks and goal-orienting. Learner-driven tasks contrast with learning driven tasks and the difference lies with who gets to decide how learning will happen. A game-informed TBLT perspective focuses on giving students agency and multiple routes of task completion with continually updated goals.

Interaction, Sykes and Reinhardt mention, is possibly the most defining aspect of games when compared with other types of media. For Sykes and Reinhardt, game-designed interaction is built on four levels of interaction. Ideational (or the interaction between learner and language, i.e., their own mind), Interpersonal (the interaction between speakers of a language) and cultural (the interaction of an individual user with the cultural expectations and values). Digital game-informed TBLT improves upon traditional applications by situating the learner into a game-oriented culture. Beyond multiplayer games, which promote both ideational language use and interpersonal, the communities that exist around games is vitally important. Using internet communities to learn, discuss and discover important information to better play the game is a fountain of possible interactions and provides the language learner with literally thousands of possible interpersonal partners.

Feedback in games is the primary way game-designers communicate to players whether or not they are progressing in the game. Actions taken by the player may lead to failure or success. Important then, for game design, is how to communicate to the player that they are failing or succeeding. For Sykes and Reinhardt feedback in games needs to be individualized, discernable and given “exactly at the moment it is needed” (p. 59). Feedback then, can be given both explicitly through the use of messages on the screen, level-ups, and tooltips. Or, it can be implicit through sound effects, or well-scaffolded tasks which teach the player step-by-step through each level of their development.

Narrative for Sykes and Reinhardt, is described as the way people transmit culture. Games generally have very well defined game-designed narratives. However, these narratives are never presented without a player and the player interaction with the game-designed narrative can lead to emergent, different narratives than perhaps the designed one. IThis is markedly different from other media genres such as books or movies. For example, game designers will intentionally create a narrative, build a world and populate it with characters and conflicts. However, unlike a book or a movie, the player then moves about that world making defined choices. Like in other genres, inside the players head there is an interpretation of the world designed by the creators. Unlike those genres however, an emergent narrative can unfold in the game-world itself, and not just in the attendant communities that surround it (such as fan fiction sites).

Motivation, finally, is described by Sykes and Reinhardt in terms of player engagement and flow. Motivation in games is not seen to reside within the individual player, but as an emerging factor in the interaction between the game-design and the player. Motivation in this view, then, is dynamic and continually negotiated and not simply an intrinsic property of the player or game. How the game uses the other four factors listed, it can help or hinder the motivation that the player brings to it themselves. The interactive nature of games can easily lead to a state of flow, or extended focus on the present moment. Flow is something of a mystical feature. We’ve all had the experience of being engrossed in a particularly good book, movie, conversation or other event. This in-the-momentness can lead to extensive exposure to whatever the game is exposing.

Conclusion

So, how can games be leveraged then? Do we just play games in English/Spanish/Korean and assume these principles will play out and, viola, language learned? Of course not. Though, then again, maybe. Some(1) researchers(2) have found(3) that just playing games leads inherently to learning outcomes, and this can be language learning in some cases.

But certainly, with the help of a Vygostkian helper, we can bring our language learners along faster. Reinhardt and Sykes help us with this and developed what they call a “bridging activities” cycle (example). In these cycles teachers help students explore these game worlds, analysis them and then perform them back in their authentic contexts.