ELFA project

“I am going to looks like stupid”: language commentary & correction in spoken ELF


Enjoy it while it lasts, folks. Fall is just around the corner…
© Nina Valtavirta

When I introduced the PhD thesis of ELFA project member Niina Hynninen (read the intro here), I outlined some considerations for studying language regulation when English is spoken as a lingua franca (ELF). The norms of acceptable English in ELF settings are not self-evident – certainly the norms of “correctness” in relation to native-speaker standards are present, but the range of acceptability might be broader that this. The answer must be found in ELF interaction itself. What do ELF speakers actually do in their real-time negotiation of “living norms”?

In the introductory post, I reviewed Niina’s interactive ELF data drawn from academic study events and the range of interactive features she examined as expressions of language regulation. In this post, I go deeper into her data and findings on four areas of language regulation. First, what kind of overt comments do ELF users make on the quality of their/others’ English? Second, what kind of explicit corrections are made in their talk? Then I move on to more subtle forms of language regulation: instances where reformulations are embedded in a second speaker’s utterance (“embedded repairs”) and reformulations involving third-party intervention (“mediation”).

Commenting on English

In the 20 hours of interactive data that Niina analysed, it was rare to find comments on another speaker’s English. It was typically students who commented on their own language. Most of this commentary was found in the student group work events as expressions of insecurity: “I’m not that good in English“, “I don’t know if the word is correct in English“, “my English is not very good I know“. Despite this uncertainty, Niina found that these comments on language quality were not accompanied by signs of communication trouble or misunderstanding. Instead, they seemed to serve as disclaimers although there were no signals of unacceptability from other participants.

One of the interesting aspects of the data was that a single English native speaker was present in each of the study events as a minority among non-native English speakers. Sometimes the language commentary of ELF speakers drew attention to this fact. In the following example, an American student (NS3) has just asked a native speaker of Spanish (S5) if he would like to speak first in their group presentation (SS = several speakers, @ = laughter):

S5: no no me because i (now) actually i can remember that you speak english a little bit better than me yeah and if i speak after you i am going to looks like stupid okay

SS: @@

S5: i think it’s better if i speak before

NS3: @okay@

<pause, 14 sec.>

S5: no but if you prefer

NS3: no it doesn’t matter @yeah@

S5: thank you very much

(Hynninen 2013: 108-9, emphasis in original)

Niina notes that in none of the study events did an English native speaker automatically assume a role of “language expert”. In the case of the American student above, he seemed to make an effort to fit in with the majority of ELF speakers – he didn’t correct or comment on others’ language, and he even adopted some of the non-standard forms that were used by other group members, most notably the phrase “study case” instead of “case study”. At one point, when he found an error in his own presentation slides, the group laughed and he responded by saying “see even I make mistakes you know @@“.

Corrections to language

In addition to commentary, language regulation can occur in the form of corrections, when a second speaker changes a linguistic detail in the previous speaker’s turn. These can be outright corrections by another speaker, but they can also be “self-initiated”, when a speaker signals uncertainty and initiates the second speaker’s correction or clarification. Overall, Niina found that “in the study events, it was generally not acceptable to correct someone’s speech and it was generally not necessary to do linguistic corrections in order to achieve mutual understanding” (Hynninen 2013: 91).

The number and types of language corrections found in the interactive data
Source: Hynninen 2013: 90

Niina counted and classified the corrections that did occur and found they were uncommon, with 33 corrections made in 20 hours of data, or one correction each 36 minutes on average. Corrections to pronunciation and grammar were especially uncommon (see the table above). 20 of the 33 corrections dealt with lexical items, especially field-specific lexis, in which teachers and mentors took on the role of language expert. Niina found that in general it was the non-native ELF users who made corrections to each other. The English native speakers in each event let non-standard forms pass uncorrected, suggesting that the boundaries of acceptability in ELF are broader than “correctness”.

Imitating the native speaker?

While language commentary and correction are overt forms of language regulation, there are also more subtle types to be found. One of these is embedded repair, when an item in the previous speaker’s turn is modified by the following speaker without explicitly mentioning it. The reformulation is thus “embedded” in a turn without any side sequence in which the correction is addressed directly. This type of language regulation shows an interlocutor’s reaction without an outright correction being made. Consider this example of an embedded repair ([ ] = overlapping speech):

S3: we can woke up at six

SS: [@@]

S1: [no thank] [you]

NS5: [we] can get up at five the way i normally do

(Hynninen 2013: 123, emphasis in original)

This example demonstrates what might be considered an expected pattern of L1-L2 talk, especially in classroom settings, where an English native speaker (NS5) reformulates the non-standard utterance of a non-native ELF user (S3). However, this pattern of embedded repair toward native speaker conventions was not the norm in Niina’s data. In many cases, the modification diverged from an English native-speaker preference:

S2: i’m sorry i’m very slow with this

S1: yeah it’s @okay@ i was slowly too @@

(Hynninen 2013: 124, emphasis in original)

There were also instances in which a non-native ELF speaker reformulated the preceding utterance of an English native speaker. In an example of this, a Canadian English speaker used the phrase “they’re on paperwork“. Then in the following utterance, a native speaker of Finnish reformulated this as “yeah it is on the paper“.

How to interpret these examples? On one hand, Niina points out that one might interpret them as examples of “deficient command of English”. After all, the embedded repairs that deviated from English native-speaker norms were all performed by non-native ELF speakers. However, the non-native ELF speakers also performed embedded repairs toward native-speaker norms as well. Niina argues that these fluctuations between norms can be interpreted as “relaxed normativity” that is typical of ELF. Moreover, these fluctuations provide further evidence that the scope of acceptability in ELF is broader than English native-speaker norms alone.

Mediation: regulating understanding

Finally, another type of reformulation is mediation, which Niina defines as “third-person intervention, a form of speaking for another where a third person mediates between two other speakers” (Hynninen 2013: 130). She has also published an independent article on this topic in the Journal of Pragmatics (Hynninen 2011). Mediation follows a set pattern of exchanges:

  1. Initial turn by speaker A, which contains a trouble source
  2. Indication of trouble or misunderstanding by speaker B (optional)
  3. Rephrasing of speaker A’s turn by speaker C (third-party mediation)
  4. reaction and response from speaker B

(adapted from Hynninen 2013: 131)

This pattern is similar to a correction or “other-repair”, but in the form of a reformulation by a third party. As Niina points out, it is language regulation toward mutual understanding. The following example illustrates this sequence in action (T1 is a teacher):

S7: mhm in reference to fire (suppression) there are some kind of organisation like <SIC> firemens </SIC> or er forest people forest prevent or fight with fire

S3: sorry (i’m) [(i don’t understand)]

T1: [er well] what the speaker would like to know is er we- no not the speaker but the er your fellow student would like to know is that is there an organisation <S3> mhm </S3> er or a system that that er that is operational in the sudan for fire suppression are there guards or are there watch towers or or what kind of mechanisms are there in place for fire suppression

(Hynninen 2013: 133, emphasis in original)

Here a student (S7) asks a somewhat disjointed question to the student presenter (S3), who doesn’t understand the question. A teacher (T1) takes the floor and provides several reformulations of the question. After this third-party intervention, the presenter (S3) re-takes the floor and the initial exchange can continue. Niina mentions that mediation was primarily performed by the teachers in this teacher-led course; it appears to function as an enabling strategy so other speakers can fully participate in the interaction.

From what speakers do to what speakers think

One of the novel aspects of Niina’s study is that she combines this interactive data with interview data from the same participants. This has enabled her to cross-analyse speakers’ behaviors with their beliefs about what is acceptable English in lingua franca settings. In the final post of this series, I’ll review these interview findings and their relationship to the interactive findings discussed above.

See also:

Part 1: Language regulation in academic ELF interaction

Part 3: Who’s in charge of English? Uses and descriptions of ELF


Hynninen, Niina (2011) The practice of ‘mediation’ in English as a lingua franca interaction. Journal of Pragmatics, 43 (4), 965-977. DOI: 10.1016/j.pragma.2010.07.034

Hynninen, Niina (2013) Language Regulation in English as a Lingua Franca: Exploring language-regulatory practices in academic spoken discourse. Doctoral dissertation, University of Helsinki. Available online: http://urn.fi/URN:ISBN:978-952-10-8639-7