Friday, March 6, 2015

Foucault’s The Order of Things: A Summary and Critique

This was one of Foucault’s major works:
Foucault, Michel. 1966. Les mots et les choses [Words and Things]. Gallimard, Paris. = Foucault, Michel. 1973. The Order of Things: An Archaeology of the Human Sciences (trans. Alan Sheridan). Vintage, New York. 387 p.; and Foucault, Michel. 2002. The Order of Things: An Archaeology of the Human Sciences. Routledge, London. 422 p.
Many would argue that this book strongly uses Structuralist ideas, although at the same time pointing to the limits of Structuralism. Indeed, many in France regarded the work as a tour de
force of Structuralism when it was published (Flynn 2005: 32).

In The Order of Things, Foucault is concerned with epistemes: an episteme is a set of ordered but unconscious ideas that are foundational in determining what is regarded as accepted knowledge in particular periods and times. This was also called the “historical a priori” by Foucault (Flynn 2005: 31; Merquior 1991: 36). But the episteme is not a general body of known knowledge or natural science. An episteme is a kind of unspoken and unconscious stratum underlying and being the precondition for accepted knowledge in each historical period, so that Foucault thought that to unearth the episteme of a period one engages in a metaphorical “archaeology” (Merquior 1991: 36).

Each episteme from one period to the next is supposed to be discontinuous and incommensurable in the sense of being radically different (Merquior 1991: 37), and there is only one underlying episteme for each historical period:
“In any given culture and at any given moment, there is always only one episteme that defines the conditions of possibility of all knowledge, whether expressed in a theory or silently invested in a practice.” (Foucault 1970: 178).
Some examples of historical epistemes as identified by Foucault are as follows:
(1) the pre-Classical period (the later Middle ages and Renaissance up to the mid-17th century), in which people thought in terms of similitudes, resemblances and antipathies (Merquior 1991: 45);

(2) the “Classical” period (the mid-17th to 18th centuries), in which the prevailing episteme stressed representation, mathesis (a science of order and measurement), and taxinomia (science of classification) (Merquior 1991: 46);

(3) the “Modern” period (the 19th century to about the 1950s), in which deep, dynamic historical explanations became important (Merquior 1991: 51);

(4) the contemporary age (from the 1950s onwards) (Merquior 1991: 37, 39).
As an aside, strangely missing in Foucault’s scheme is the ancient Graeco-Roman world, but let us not dwell on this point.

When one episteme is replaced by another, this process is by some irrational and cultural revolution akin to religious conversion or mass psychology, not by rational argument or evidence (Merquior 1991: 37).

As Foucault’s fierce critic José Merquior argues, all these central ideas of Foucault are highly questionable.

Foucault’s insistence that each episteme represents a sharp break with the previous one and a fundamental discontinuity does not withstand much scrutiny (Merquior 1991: 58). Merquior sees deep problems in Foucault’s attempt to classify Western intellectual history into his four periods with underlying epistemes, and identifies devastating problems as follows:
(1) the transepistemic objection: some fundamental ideas and epistemic principles which span more than one era (Merquior 1991: 62–65);

(2) “epistemic lag” objection: when significant ideas and theories from a previous episteme persist into the next period (Merquior 1991: 65);

(3) the return of earlier ideas or systems of thought: when ideas or theories supposedly overthrown by a new episteme actually reappear and are influential (Merquior 1991: 65–67);

(4) heterogeneous epistemological ideas within each of Foucault’s specific periods: often we find Foucault’s characterisation of the reigning ideas of each period is shoddy and simplistic and does not do justice to the actual state of affairs in that era (Merquior 1991: 67–69);

(5) instances when some fundamental theory supposedly characterising each episteme collapsed well before the end of its period (Merquior 1991: 69).
To give some examples, the ideas supposedly at the heart of the Classical era were already clearly in evidence in the pre-Classical period, especially in the realm of the emerging science of cosmology (Merquior 1991: 58). Taxonomy was already a fundamental part of the Aristotelian philosophy that was very influential in the Middle Ages (Merquior 1991: 59).

Within alleged periods where one episteme supposedly ruled, one can find a great degree of pluralism in which it is difficult to see how some monolithic underlying episteme could really have existed (Merquior 1991: 62–68).

And finally we get to a really devastating problem with Foucault’s periods with underlying epistemes: Foucault was not concerned with the question of whether ideas in epistemes or in general knowledge overarching them could be justified as rational or objectively true. He had no interest in The Order of Things in whether there was a development of growing ideas that can rightly be regarded as objectively true (Merquior 1991: 71). But in any history of human knowledge or the epistemological ideas underlying the sciences this is a fundamentally important question.

By contrast, in Foucault’s The Order of Things there is a lazy and unproven assumption that what Foucault called the human sciences (namely, the biological, psychological, and social sciences) and the principles of any given episteme cannot get to objective truth or be rationally defended as against other epistemic principles or ideas.

If you are engaged in the history of knowledge as Foucault was, but are not willing to seriously address the question whether empirical truths and foundational epistemic principles can be rationally defended or are objectively true, then the whole project has missed the point.

Gutting, Gary. 2003 (rev. 2013). “Michel Foucault,” Stanford Encyclopedia of Philosophy

Kelly, Mark. “Michel Foucault (1926–1984),” Internet Encyclopedia of Philosophy

“Michel Foucault,” Wikipedia

Merquior, José Guilherme. 1991. Foucault (2nd edn.). Fontana, London.

Flynn, Thomas. 2005. “Foucault’s Mapping of History,” Gary Gutting (ed.), The Cambridge Companion to Foucault (2nd edn.). Cambridge University Press, Cambridge, U.K. and New York. 29–48.

Foucault, Michel. 1970. The Order of Things: An Archaeology of the Human Sciences. Vintage Books, New York.


  1. Foucault abandoned these structuralist ideas in the 1970s.

    Of course, Post-Keynesians generally adhere to a more modern version of Foucault's theory in the work of Philip Mirowski.

    "By contrast, in Foucault’s The Order of Things there is a lazy and unproven assumption THAT THE SCIENCES and the principles of any given episteme cannot get to objective truth or be rationally defended as against other epistemic principles or ideas."

    Em... you might refer to the subtitle of the book that I can only assume you haven't read:

    "An Archaeology of the HUMAN SCIENCES"

    I don't know how many times I have to point this out to you before you allow the distinction to sink in. I suspect that you are simply not listening because you don't want to.

    1. By the phrase "the sciences" I did mean both the relevant natural and social sciences that Foucault was referring to by the expression "human sciences": that is, the biological, psychological, social sciences.

    2. I have fixed the sentence to make it clear.

  2. We might also note that the critic that LK is drawing on is a well known neoliberal.

    1. Well, Phil, why should that be a reason to reject his critique of Foucault?

      This is a shoddy ad hominem fallacy: it is like the libertarian who implies that you can't believe Keynes because Keynes was a supporter of eugenics.

  3. This book has it's flaws, but I don't think Merquior's arguments work.

    It's important to remember Foucault did a lot of interviews where he made clarifications he should've made in his books. I recommend reading all of those before grappling with a Foucault book.

    Addressing Merquior's criticisms:

    1. An 'episteme' doesn't deal with all fundamental ideas and epistemic principles. It's a way of how we think about SOME things - it's not a comprehensive way of thinking about everything.

    2. The persistence of ideas and theories into another episteme doesn't refute the notion of epistemes. Not everything needs to change for an epistemic shift to occur.

    3. There's no reason, based on Foucault's theory, why this couldn't occur.

    4. (I'd have read the article to make sense of this one. My general reaction is that Foucault isn't trying to make sense of all the heterogeneity, just explain things on the basic/epistemic level, and that the heterogeneity would depend more on various political issues and not the general/epistemic way of thinking at the time).

    5. First, a theory which 'characterizes' an episteme and the episteme itself are not the same thing. Second, a fundamental theory (as in fundamental to the episteme continuing to be accepted over the long haul) to an episteme would almost certainly be a prerequisite to the collapse of an episteme - it gets people thinking "Hm...if this is wrong...then that means that would have to be wrong...but if that's wrong too". How could an episteme end otherwise? Someone comes up with an idea that is immediately accepted by just about everyone totally ending the episteme in a few short months? But people didn't understand their epistemes, so it's unlikely they'd figure out how it works. More likely, they figure out a problem with one of its fundamental (or perhaps it's more precise to say 'principal') theories, and the episteme starts to unravel.

    If anything, a principal theory didn't end before the epistemic shift, it would be an argument against Foucault's theory. After all, if we can't understand our epistemes - they're so fundamental to how we think/it's so difficult for us to conceive on thinking in a new way - wouldn't it be virtually impossible for someone kill the damn thing all at once? More likely, it happens over time, with no precise end/start date (only an approximate one, something Foucault is explicit about).