In today’s Higher Education Supplement Paul Frijters and Tony Beatton have a go at this article from a couple of weeks ago by Andrew Leigh for his ‘various platitudes about the advantages of diversity’. Frijters and Beatton say (recycling what Frijters said on Andrew’s blog):
one of the biggest dangers for a scientist is to get sucked into simply stating things that go along with what is commonly believed but are not really self-evident at all. It’s very easy to make statements that go with the grain without applying the same amount of thought given to statements that go against the grain.
I’m all for scepticism. One of my most useful lessons in social science came when, as an undergraduate in the late 1980s, I set out to write a critique of a chapter on the ‘New Right’ in book by one of my teachers. I decided to fact check every statement made, and was very surprised to find that quite a few things were simply wrong and many claims were not supported by the evidence I could find. Academics as much as any of us are least sceptical when confirming their existing prejudices for or against an idea, and as a result most vulnerable to critics hoping to take down their arguments.
But it’s hard to see what Andrew has done wrong here. These were the ‘platitudes’ he offered, numbered by me with my commentary in italics:
1. From a purely economic standpoint, immigration supplements our labour market with much-needed skills.
There are immigrants who don’t have useful skills and have high rates of unemployment, but overall there is plenty of evidence that this is a correct statement.
2. In a deeper sense, immigration is valuable because it weaves new threads into our cultural tapestry. Native-born children have much to learn from their migrant peers, just as adults can gain a deeper understanding of the world from yarning over the back fence with their foreign-born neighbours.
It’s hard to quantify this, and it is not much of an argument for immigration in itself, but certainly talking to people with very different life experiences can be interesting.
3. And our restaurants would be bland imitations of themselves without the flavours brought by successive waves of Italian, Thai and Vietnamese immigrants.
There was a debate about this at Andrew L’s blog – while as Frijters suggests they may have displaced 100 different ways of cooking parsnip, it seems wildly implausible to believe that we are worse off for it. Nobody forces us to eat at ethnic restaurants rather than have a different parsnip concoction each day.
So I don’t think there is anything much wrong with these statements. They are just not substantiated in the way that the rest of what Andrew has to say is backed up, and you would not base an immigration program on them. But he is not saying we should. As I think he has noted himself somewhere, they serve a rhetorical purpose in a short article.
If, as a social scientist you are trying to argue something that goes against your audience’s intuitions, it can be useful to say that you don’t reject all their worldview, because otherwise they could just dismiss you as a right-winger/left-winger that they don’t need to listen to. It’s not enough to be right, you need to convince other people that you are right as well.
There is also the ‘yes but’ problem, which I have often faced. In the confines of the word limit an author sets out to argue one proposition (in Andrew’s case, that linguistic diversity is associated with lower trust). Often this will be quite convincing – why else would you have written it? – so people who disagree with your conclusions resort to ‘yes but’ objections, ignoring what you have said in favour of other arguments against your conclusion. It is useful to try to take these off the table and keep the focus on your main argument.
So while I am all for fact checking and robust debate (indeed, I was critical of one aspect of Andrew’s argument myself), I don’t think there is anything wrong with using substantiable though not substantiated assertions in an article aimed at a general audience. Indeed, it would be hard for social scientists to communicate to the general public without doing this sometimes. And would the space they leave empty because they cannot prove every point in 750 words be filled by people with more respect for facts? I doubt it.
13 thoughts on “When is it OK to make unsubstantiated assertions?”
Andrew, thanks for an interesting post. I expect that most readers will be focusing on the second part of the Frijters/Beatton article (where they say that we shouldn’t blame policymakers for Indigenous health outcomes) more than the first.
What I found most interesting when I posted my original piece on my blog was Harry Clarke’s comment. He said:
“Fair enough Andrew but would you expect an increase in diversity (and in particular an increase in the range of languages spoken) to increase trust. If you can
I’m reminded of Hayek’s arguments about ‘contructive rationalism’ (see LLL vol. 1). Andrew L made several points that most people would think, at a gut-intitive level, are correct. Now Paul Frijters argues that intellectuals should be sceptical and question more. Yes, they should. But as Andrew N indicates we can’t go around deriving all our arguments from first principles all the time. For some purposes, we need to accept (some) things as given and work from there.
Andrew (Leigh) if I cannot speak to someone (because they speak a different language) can I ever trust them more? Could it be that introducing the possibility of speaking to someone in their own language reduces trust . My assumption is ‘no’.
Andrew Norton (on another issue, sorry) – its slow for me to get onto your blog much of the time.
“its slow for me to get onto your blog much of the time.”
For me too, Harry. I was thinking it was my computer or ISP.
Actually Andrew, sorry to get off topic, but its not just slow for me — I am getting rejected sometimes also.
Guys, I have complained twice to Yahoo (the host) about slowness/unavailability but they say it is working ok – which it no doubt was when they checked.
This url often works when the normal one does not:
Nobody can even give me a theory of what might be wrong, so I am reluctant to go to the hassle of changing from Yahoo when they may not be the cause.
“Andrew (Leigh) if I cannot speak to someone (because they speak a different language) can I ever trust them more? Could it be that introducing the possibility of speaking to someone in their own language reduces trust . My assumption is
I suspect there is a limit on the number of times the site can be accessed. So I have cleared out my history file and then been able to access the site. Also when using the other home computer I’m able to access the site. Rule of thumb – you can come here twice in a day before hitting problems and having to muck around with alternate addresses or coming via different links from other blogs.
Andrew – I too find the site very slow and often impossible to load.
I’d be willing to bet, without any research at all, that 100 years ago in england there was the one main method of cooking a parsnip that survives until today over there. Boil the crap out of it.
Yeah, I read that article on the weekend and – at least wrt their comments on Andrew L’s article – got the impression that these guys were being a bit petty . I’m sure they could have found much better examples to illustrate their point. Professional jealousy?
Andrew Norton makes a very pursuasive case for being lenient towards what one can expect from social scientists in a 750 word article, and I completely agree with his conclusion that ‘I don
Paul – I don’t think I ever lie, but I won’t necessarily raise all the issues if I think they will just cause the debate to be distracted from what I want to talk about and persuade others of.
Or I will take things as given for the sake of a discussion. For example, I don’t find most forms of inequality to be of any inherent concern. But others do, so I am willing to discuss the implications for inequality of things I support, such as market reform of higher education.