Analysing learners’ language: a method to their madness

IMG_1995A lot of effort in language teacher education goes into making teachers aware that learners of a second language make errors that are not random or unprincipled, or indeed simply transfer from the first language, but rather obey a certain logic of their own.  Learners are not simply defective speakers of the target language, but are constructing their own interlanguage (Selinker, 1972), also referred to as learner language.  This means that the developmental stages of their language acquisition should be treated as necessary steps in the language learning process, rather than as strings of errors requiring remediation.

This idea has implications for the planning of language teaching, since it suggests that teaching should be tailored to learners’ in-built syllabus, rather than organised according to a separate, grammar-based syllabus.  If we accept the interactional hypothesis, for example, which claims that language is acquired through using the target language in the negotiation of meaning, we should allocate class time to communicative activities which engage learners in such language use, and support learners in participating in activities and reflecting on their engagement.

But the notion of learner language also has implications for the evaluation of language learning and teaching.  Teachers traditionally identify and correct learner production by underlining written errors, for example, or recasting oral errors of vocabulary, grammar or pronunciation.  But if we want to better understand learner language, more analytical methods are required.

Analysing errors: language accuracy

A CARLA resource on learner language explains that

“An error analysis – and teacher corrections – should ignore unsystematic performance slips (mistakes) and focus on errors, which are systematic violations of the rules to which the learners have been exposed; these tell us something about the learner’s current knowledge of the rules of the language being learned.”

Dyson (2010) distinguishes three different ways of studying errors (1):

  1. error analysis
    1. collect a sample
    2. identify grammatical errors
    3. record error frequencies
    4. explain errors
    5. repeat procedure with other morphemes
  2. obligatory occasion analysis
    1. select a morpheme
    2. identify and count obligatory occasions
    3. count suppliance of morphemes
    4. calculate accurate use as a percentage
    5. order devices implicationally in relation to other morphemes
  3. frequency analysis
    1. select a linguistic variable
    2. divide data into equal periods
    3. identify different devices used
    4. calculate frequencies
    5. identify dominant device at each point in time

More advice on error analysis is provided in the CARLA resource.  However, we should remember that accuracy is only one dimension of learner language: the complexity of a learner’s production is another important aspect.

Assessing complexity

Tarone and Swierzbin (2009) suggest measuring syntactic complexity by counting any of the following:

  • the number of sentences containing more than one verb
  • use of complex noun phrases
  • number of verb arguments
  • types of dependent clauses

For lexical complexity, variety can be analysed using type-token ratio: “the total number of different words (types) divided by the total number of words (tokens) in a given segment.” CARLA resource.


The fluency of a learner’s spoken production is perhaps harder to evaluate.  Different measures include

  • Breakdown fluency (e.g., time filled with speech, no. of pauses, filled pauses)
  • Speed fluency (e.g., speech rate measured as words per minute, syllables per minute)
  • Repair fluency (e.g. false starts, repetitions)

(See Skehan, 2009, cited in De Jong & Hulstijn, 2009)

Learner language can therefore be analysed in a number of ways, looking at its complexity, accuracy and fluency. We can look at several different learners, under different conditions of language production, or one learner at different points in time.  These measures are different from the kind of evaluations teachers constantly conduct, both formally and informally.  But they provide a different picture of learner language development, allowing teachers to take a step back from day-to-days concerns to see how their learners are really doing.

1. Another approach advocated by Dyson (2010) involves emergence analysis, based on Pienemann’s processability theory.  This approach is no doubt too complex to be useful for practitioner research.


Overview of Complexity of Learner Language

Overview on error analysis.

De Jong, N., & Hulstijn, J. (2009). Relating ratings of fluency to temporal and lexical aspects of speech. EALTA 2009 [slides]
Dyson, B. (2010). Learner language analytic methods and pedagogical implications.
Australian Review of Applied Linguistics, 33(3), 1-21. [PDF]
Selinker, L. (1972). Interlanguage. International Review of Applied Linguistics, 10, 209-241.
Skehan, P. (2009). Modelling second language performance: Integrating complexity, accuracy, fluency, and lexis. Applied Linguistics, 30(4), 510-532. doi: 10.1093/applin/amp047
Tarone, E. & Swierzbin, B. (2009). Exploring Learner Language. Oxford: Oxford University Press.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s