I have become aware of what Yuval Noah Harari calls the "Historical Knowledge Paradox", in which the utility of knowing the past is placed in conflict with itself. Can somebody please help me to see how this supposed paradox fits within the philosophical discipline of epistemology?

See page 67 of Yuval Noah Harari's Homo Deus :

"This is the paradox of historical knowledge. Knowledge that does not change behaviour is useless. But knowledge that changes behaviour quickly loses its relevance. The more data we have and the better we understand history, the faster history alters its course, and the faster our knowledge becomes outdated."

  • 1
    Bit vague. Can you shape this into a proper question, rather than just a request for opinions? You might find this perspective on the evolution of a co munities 'metis' or practical-craftiness in how to live, and problems with legibility of this between generations and eras: aeon.co/ideas/…
    – CriglCragl
    Apr 5, 2018 at 12:14
  • As you can see by the current close votes, adding a bit of context and own thoughts would be more helpful than throwing in a quote.
    – Philip Klöcking
    Apr 5, 2018 at 12:58
  • "The more data we have and the better we understand history, the faster history alters its course...". It is always good to study history and maintain the record. However, it is only partially true that more data makes change. And this is why the Frankfurt School adopted Freud. The "frozen history" of the superego, backed by the authority and order of a successful Oedipus, can block our ability to act on new data. The way to move forward is to loosen the hold of the super-ego. And yes, in times of crisis this can happen; it is one way it can happen.
    – Gordon
    Apr 5, 2018 at 13:48
  • I should mention that reason is the proper evaluation of father (the frozen attitude, backed by authoritarian overtones), and this evaluation of the old man takes place in the adult Ego and enables us to act on facts without fear of "Father"; to evaluate facts in the light of reason which takes account of the facts without fear, and acts on its findings if appropriate.
    – Gordon
    Apr 5, 2018 at 14:24
  • I think Bayes is better at handling new data than Freud. Freud kills "father" but Bayes renames him "prior". Apr 5, 2018 at 14:32

3 Answers 3


There are several takes of this throughout the history of philosophy.

One of the earliest examples I can think of is the philosophy of Dilthey/Misch, where the Kantian idea of transcendental [German transzendentale and transcendental also in non-adverbial form, i.e. necessary conditions of the possibility of experience, not the same as transzendente and transcendent in non-adverbial form, i.e. beyond all possible experience] logic is relativized historically, i.e. the factual and embodied knowledge of various times is understood as founded on factual and logical/conceptual conditions of that time. Both can change, which is exactly what makes history possible and actual reality. A related concept evolving around that time and being coined shortly thereafter out of that idea is the hermeneutical circle.

The most famous example is by Michel Foucault, the historical a priori or episteme, as laid out in The Archeology of Knowledge and The Order of Things respectively. It expresses a similar thought to the one of Dilthey/Misch.

What both takes have in common is that epistemologically, they analyse the factual and conceptual historical conditions and show how certain insights have been dependent on them and both necessary for the time and necessarily true only as embedded into that particular historical context.

Coming to the paradox: As I see it, it confuses two aspects that above-mentioned authors are careful to distinguish - historical process [meta-layer and abstract fact] and historical setting [factual situation at a time]. An improvement of the understanding of history does not exactly change the pace or nature of the historical process. A change of the historical setting does, i.e. coming to know more or other "facts" in general. Knowing more about history is but a small aspect of it.

In other words: It is new knowledge that's never been there (or been completely lost) that outdates knowledge, no matter whether historical or not. But historicity of knowledge is a fact that is only epistemologically accessible in hindsight. It is only what already has happened that we can know, i.e. epistemologically accessible. These are the huge insights of Dilthey, Misch, and Foucault [well, and others]. Therefore, saying something about the current state of the process of history at large is indeed paradoxical as it is impossible by the very nature of historicity. All we can say is that everything that is and ever has been (and ever will be) is only real insofar as it is historical.


Harari seems a bit carried away by the impacts of rapid change, which is to project the current historical context on to all times, just like say Malthus or many others saw a contemporary situation as key to the future. Foucault identifies the development of knowledge with power structures. These may be largely invisible during times of peace, but absolutely limit the pace of real change, and of it's 'self awareness'. Study of history affects government policies very little, compared to political tribalism.

Knowledge with power remains, when there is a conflict with knowledge with less power - or that is less integrated with structures previously succesful in the conflict of ideas and organisations.

Many ideas, like what human nature is, change generationally, are renewed by stories and events, but only really manifest deep change in moments of crisis. It just seems misguided to think we can know which knowledge will change history, even which knowledge of history.

Stories that sway us have power, and if they are forgetten or there is not time for us to absorb them, they lose it. Conscious change happens when we create stories out of knowledge of audience, to create change that is percieved as needed. Only time will reveal which win, leaving us try and u derstand audience, needs, and shape how our own spirit partakes in this zeitgeist.

  • Thank you. Can you elaborate a further "rapid change" part. It seems you are up to something interesting.
    – KwanzaKymi
    Apr 6, 2018 at 13:08
  • @KwanzaKymi: Knowledge doesn't change history, power and stories do. We get stuck in paradigms of our era, until something breaks us out of them
    – CriglCragl
    Apr 6, 2018 at 15:22
  • 1
    Thank you. Your comment increased my understanding of what you were meaning.
    – KwanzaKymi
    Apr 8, 2018 at 8:20

This seems to be similar to the Efficient Market Hypothesis, which states that all information is incorporated into stock prices, and thus learning information does not help one pick stocks. For information to be useful, it would have to either no be incorporated into prices (e.g. insider knowledge), or be something that remains useful even when others are acting on it (for instance, diversifying one's portfolio is useful even if others are doing it).

There is a bit of a fallacy in saying "But knowledge that changes behaviour quickly loses its relevance." Knowledge that changes one person's behavior (may) lose (some of its) relevance for other people. If one person learns a piece of knowledge, that person can still benefit from it. It's only when everyone learns it that no one has an advantage. Thus, knowledge can be put into three categories: knowledge that has not been fully exploited by others, knowledge that remains useful even when exploited by others, and knowledge that has lost its relevance through exploitation by others. Only some knowledge falls in that last category.

  • Tangentially, or maybe in parallel, Christians have an specific advantage in competitive real estate markets! Whenever a house is offered in which somebody died, home-shoppers get freaked out and won't bid on it. So somebody who has confidence in knowledge about the afterlife will get a bargain! Apr 5, 2018 at 21:31
  • uaoooo! You are correct. I did not think about Efficient Market Hypothesis but it seems correct to me. Thank you.
    – KwanzaKymi
    Apr 6, 2018 at 13:07

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .