Hacker News Reader

Comment by Garlef | Hacker News Reader

Mediocre/bad article. It's mostly a rant against decolonisation in general.

The actual treatment of why this mostly does not apply to maths is very thin.

And there actually is a problem with the notion of mathematical truth: Not within mathematics itself but in its transfer to sociological questions. Here, simply presupposing that concepts can be defined with a universal meaning and statements have a universal truth is simply not true. Mathematics is not advanced enough to discuss these matters properly. (So: New notions need to developed)

Since most readers here have a background in programming, here's an analogy of what the issues are: Sociology is about distributed systems. And these are more complicated. For example thinking about time in distributed systems requires quite a lot of non-trivial tech (vector clocks, etc). It's even more complicated for sociological notions.

But in general, I agree: The topic mostly does not apply to maths.

View on HN

I am not sure if I agree. This is a debate that isn't really related to diversity since people have asked things like how mathematics became so successful in other sciences like physics.

I guess you could apply that questioning to sociology. You could ask why some groups are more prone to crime than others using a statistical test. You could maybe use data to try to reason about sociological questions. It might turn up some ugly results. Or maybe people could ask why standardized scores differ between groups and instead of concluding there must be a difference in the test takers, they conclude the mathematics on the test must be at fault. So where it is convenient people question the validity of mathematics itself.

But on the other hand when someone tries to quantify diversity statistics and says there are too many white men who are CEOs or there is a black-white wealth gap, it seems like mathematics is perfectly legitimate. Should people watching the populations of white people be reminded that their statistical techniques are incorrect because they haven't been 'decolonised'?

Of course it's a very deep topic in the philosophy of science, foundations of math and logic, and philosophy of mathematics to ask about what math really is and where it applies but that's not what is being done here when there is talk of decolonisation.

The point of sociology, like all the humanist sciences (even languages) is to change the world, not to study it. The highest good in social sciences is social change, and methods to achieve it. You're applying positivist ideas to non-positivist sciences.

For humanist sciences truth is a tool in service of a higher, moral goal. Just like lying is another tool. Both are to be used, and whether some claim is true or false is not nearly as relevant as whether it can change society. For languages truth is irrelevant, after all you're studying fiction. Fiction is mean to convey "higher truths", which can in practice be anything from amusement to outright deception.

Often people in these sciences go further and go for outright postmodernism, which disputes that truth and falsehood even exist at all. It's just a matter of perspective.

that sounds terrifying. they are given grant money for this.

Something being terrifying does not make it true.

This is a great example of precisely what I find problematic about applying logic in a simplified way.

In itself the argument seems valid. But it relies on assumptions that are (at best) highly debatable:

First, the author of the post assumes that the definition of the humanities he presents is universally true (carried by everyone). It is not: I never heard of it before, and I don't agree with it. To me it's just part of a taxonomy of how departments in universities are partitioned.

Second, even under the assumption that this definition was universally shared, the argument presupposes that this definition would then also fully determine the actual praxis. Which I highly doubt.

Third, the point about lying is just a bunch of fear mongering BS.

Etc etc.

> mathematics itself but in its transfer to sociological questions > Sociology is about distributed systems.

You make the argument that sociology is its own thing that can't be conflated with math which I think every one agrees with, isn't that the premise of the article?

Sociology has its time and place and thats in sociology, just like Math has its own time and place Math.

Yes and no - The underlying issue that prompted me to bring up the lack of adequate mathematical tools to discuss sociological notions is the following:

Mathematical truth and logic is often used as a gold standard for universal truth. And it is then often used as a straw argument to simply push away complex and uncomfortable questions. (A hyperbole of such a misguided line of reasoning would be "Software is just maths/logic and therefore pure and we so don't need to think about sexism and other social issues in sofware engineering.". A more nuanced version would be parts of the article we are discussing.)

But the "default tech" when thinking about mathematical truth (for example predicate logic) is in itself biased towards universally definable truths: A given predicate is universally defined a-priori.

There's no tools for things such as conflicting definitions and speakers positions in predicate logic: In reality, different actors in a discussion might not agree on what even they are talking about.

Take as an example the often discussed statement "White people can not experience racism". If you try to naively approach this with the toolset of universally definable notions you quickly realize that it can not be true (assuming a definition of racism that is not in itself skewed).

But there is a lot of things going on here: Who are these white people? Is the statement "This person is white" even based on a well defined predicate (Hint: It's not)? Etc etc.

And so: If the only logic in your toolset is predicate logic, it's hard to grasp sociological issues.

The problem is that at least in my experience, people who are trained in mathematics are generally rather aware of the limitations of models. I find that it's mostly practitioners of other disciplines (sociology, economics, sometimes also CS, etc.) who cargo cult mathematics without understanding it deeply enough.

That said, I'll agree that it sometimes feels like there's two types of people, the ones that think everything is amenable to rigorous and objective reasoning and the ones that think nothing is. But it's important to realise that both these modes of reasoning are important. For example, a mathematical proof can show that there exists no perfect voting system under some reasonable definition of "fairness"[0], but the theorem itself cannot decide whether that definition of fairness is really the one we care about.

But it's one thing to say that "there are more modes of reasoning than ones that lead to objective outcomes" and quite another to say that objective facts play no major role in our society and that getting people to understand that you can't always bend reality to your preconceptions isn't incredibly important.

[0]: https://en.m.wikipedia.org/wiki/Arrow%27s_impossibility_theo...

But should that technology even be a part of mathematics? Different fields/applications will have different needs so I would imagine that these critical discussions should be part of each field and its methods. Eg: physicists don’t go around blaming math because specific physical systems need different models.

With "mathemathical technology" I'm not taking about software but actual mathematics.

And no one is blaming maths.

If physicists realize there's something missing, they develop the maths they require. Or they ask their maths friends to do it. And it might happen if enough people care to pull it off. Same applies here.

(But it's not very likely since most sociologists are not very educated in mathematics)

Reply on news.ycombinator.com