In March 2016, Louise Greener gave a presentation to a BALEAP meeting at the LSE called "Occupying the niche: opening in-sessional to all". Twenty-three slides, arguing that Durham should stop restricting its in-sessional academic language provision to international students and open it to every student in the university. One slide carries a line from a Durham lecturer, recorded the week before: "I think of my international students as having a disability". Another records what her colleagues concluded, which was that provision should carry on in a form that did not "fuel the deficit model". The framing she chose for the change was "no-one's first language is academic English".
That sentence is hers, and we did not find it ourselves. She told us. The flattering version of this story is available and we are not going to tell it, because it would have us arriving at her position independently and then discovering she agreed, and that is not what happened. We heard it from her, in passing, as the kind of thing somebody says when it has been obvious to them for years. What we can claim for ourselves is only that it landed, and that we have not been able to put it down since.
What the line does is move the question from who a student is to what that student has already been exposed to. The thing we did not understand before we started reading for this was that learning English for everyday use and learning English for academic study are two different activities. Michael Halliday's "Towards a language-based theory of learning", in Linguistics and Education 5 (1993), separates learning language from learning through language, and a degree asks mostly for the second. That difference is roughly what separates EAP from language teaching in general. Ken Hyland (2018), in Language Teaching, defends EAP as specialist work rather than ordinary English teaching relocated to a university.
Once you hold the distinction, a British eighteen-year-old in week three of a first year is doing a version of the same task as a student who arrived in September from Shanghai. They are at very different points along it. Neither of them has been writing literature reviews since childhood.
Exposure rather than nationality
The part we would add is about class, and it is ours rather than hers, so it should be judged separately and it should not be attributed to her if it turns out to be wrong. If you come from a household where nobody has read a book, you are not going to know the same things as somebody brought up by two parents who are researchers. You will not have been brought up in the same language. The distance into academic English is longer for some students than for others, and what sets the length is prior educational exposure rather than nationality. That is the reality of it, and it is not a comfortable thing to say to a sector that files students by fee status.
We are aware that this is not our field. Basil Bernstein worked for decades on how social class shows up in the language of schooling, and Karl Maton's Knowledge and Knowers (Routledge, 2014) builds Legitimation Code Theory on that work. The crude reading of Bernstein, the one where working class children simply have worse language, is the reading Bernstein's defenders spend their time correcting. We have not read enough to make the argument in those terms and we are not going to borrow the vocabulary to sound as though we had.
The claim we are making is narrower than Bernstein's and it is about exposure. A student who has met academic prose at home, in a house that had some, starts closer to it than a student who first meets it on a reading list. We have not measured that. We believe it and we have not shown it, and those are different states.
Sorting rules, including ours
Greener's own supporting material goes further than the slogan, and it goes in an administrative direction, which is the part that matters to a company selling software to a university. Her slides state that "'Home' and 'international' are essentially financial designations", and that "'Native' and 'non-native' are fraught with complication and potentially unhelpful". Her alternative is not a slogan either. It is a proposal about how to allocate: "Think about distributing students according to academic level / level of academic acculturation rather than 'native-ness'." That is a sorting rule offered in place of another sorting rule. It says the categories a university already uses are artefacts of how the money arrives rather than descriptions of what students can do, and that a better sort exists and is available to anybody willing to use it.
A software licence is also a sorting rule. Asked what would change about the way we build the product if a university licensed it for a whole cohort, our answer was that nothing would change, because it would not be for a whole cohort. It would be for the international students inside the cohort. That is the premise the product is currently sold on.
So the position this essay has arrived at, that the thing being learned is nobody's first language and that the distance to travel is set by prior education rather than by nationality, points straight at our own commercial model and finds it inconsistent with itself. If academic English is a second language for everyone in the room, then a licence issued on the basis of fee status hands the tool to a group of students selected by something other than the thing the tool is for.
It is worth being exact about which way the money points, because we would like to claim the honourable motive here and cannot. A whole-cohort licence is a larger contract than a licence for the international students within that cohort. The money, in other words, points at the arrangement we are not proposing. What holds us to the smaller licence is that it is legible: it maps onto a budget line that already exists and onto a story about international students that the sector already tells about itself. Selling something because it is easy to explain is a worse reason than selling it because it pays better, and it is the reason we have.
What we can defend, and what we are leaving open
The claim we think we can actually defend is about time rather than capability. Students learn academic English one way or another in the ordinary course of taking a degree. They just do not do it as quickly as they could, and what a tool of this kind might do is shorten that. It is a modest claim and a testable one, and we have not tested it. Nobody has used UniFluent for a hundred hours. If it is true, it is true of a first-year historian from Sunderland in the same way that it is true of an engineering student from Shanghai. We cannot find a version of the argument in which it is true of only one of them.
We are not going to resolve the licensing question here, and we would rather say plainly that we are leaving it open than leave it open quietly. The decision has not been made inside the company, and writing a commitment onto a philosophy page in order to look consistent, before making it anywhere that costs anything, would be precisely the move this page exists to avoid.
Some of the hesitation is not cowardice. Who receives a licence is a decision a university makes and not one a supplier makes for it, and the case for funding academic language provision for British undergraduates is harder to put to a finance committee than the case for funding it for the cohort the sector's finances depend on. It took a presentation and a staff consultation to make that case at Durham, and that was from inside the university. We are a maths student and an engineering student with a piece of software. We would also expect provision opened to everybody to be taken up mostly by the students who already suspect they need it, which is a guess rather than a finding: any university that has opened its in-sessional provision has attendance figures that would settle it, and we have none.
The rest of the hesitation is that we have not yet followed our own logic anywhere it would cost us anything, and there is no flattering name for that.
Where the evidence stops
The last thing is the one that undermines most of the above. Two students have told us the product was useful. Both are Chinese, both are learning English in order to study in it, and one of them had used it for two days when she wrote. They are the students the current sales model already targets. The extension we are most attached to, the one about prior exposure rather than nationality, is the part of this essay with no user evidence behind it whatsoever. We have never handed the product to a British undergraduate who was struggling with a reading list and asked whether it helped. Until we do, "this is for everybody" is a position we hold rather than a finding we have, and anyone reading would be right to treat those differently.
Nor is the product obviously shaped for that student. The one place a student's own course reaches it is vocabulary: material they upload is used to pull out words from their own subject. The voice tutor and the writing tutor do not see any of that, and we have not built the link yet. Everything other than the vocabulary is general-purpose English, which is the part a student who has spoken English since birth needs least. So if a university said yes to a whole cohort tomorrow, we do not know whether what we have built would be much use to that student, because nothing in it was designed with them in the room.
