Does knowledge require consciousness for the entity that knows? In other words, is it the case that only conscious entities can know things? I was led to ask this question by considering whether or not the search engine Google knows things. It seems true to say that Google knows many facts, but does it really? After all, Google, or even the Internet more generally, does not (as far as we humans know!) have consciousness. So, my question is, can only conscious entities know things?

  • If one can know something without intentionality -- without reflecting on said knowledge -- then perhaps consciousness is optional.
    – Michael
    Apr 5, 2022 at 7:38
  • On a related note, does knowledge exist when an external cue or trigger is required for recall? Language transformers such as GPT-NeoX (third-party playground available here) can produce intelligent output based on a user-provided prompt. If such a prompt is necessary, is the output nevertheless an expression of knowledge?
    – Michael
    Apr 5, 2022 at 8:11
  • Does this answer your question? Do machine learning algorithms have knowledge (if not justified true beliefs)?
    – Dave
    Apr 5, 2022 at 14:50
  • No. Knowledge is an object, and objects don't have needs. On the contrary, you, you need not only consciousness, but a number of other resources in order to get knowledge.
    – RodolfoAP
    Apr 7, 2022 at 10:18

6 Answers 6


Short Answer

Depends. Many philosophers recognize different types of knowledge. Gilbert Ryle, for instance, in his seminal work The Concept of Mind recognizes knowledge-how (SEP), for instance, which is commonly understood as a skill. It is possible, therefore, to have skills and not be consciously aware of them. According to a naturalized epistemology, many aspects of knowledge have been demonstrated to be subconscious.

Long Answer

The philosopher Robert Audi posited that one of prime sources of knowledge is consciousness, and certainly conscious knowledge-that is often held up as the quintessence of knowledge. Knowledge-that requires language skills and demonstrates self-reflection, and is the core content of what lies between the covers of philosophy texts. But there are situations where a philosopher is pressed to concede that consciousness knowledge-that isn't comprehensive.

Let's start with a sleepwalker. What should a philosopher make of the fact that a person who is in an altered state of consciousness might function to an uncanny degree as a conscious person? Somnambulists (Mayo Clinic) may engage in relatively complex behaviors:

Do routine activities, such as getting dressed, talking or eating
Leave the house
Drive a car
Engage in unusual behavior, such as urinating in a closet
Engage in sexual activity without awareness
Get injured, for example, by falling down the stairs or jumping out a window
Become violent during the period of brief confusion immediately after waking or, occasionally, during sleepwalking

I don't know about you, but driving a car while talking to someone in sleep seems to be a perfect case to demonstrate knowledge-how and knowledge-that are both possible with out consciousness. If a philosopher of mind pays attention to psychology, then it's rather certain that consciousness itself is but a very limited aspect of the human mind, and that intuition and habit are far more weighty in terms of human behavior. From abnormal psychology and its exotic cases that demonstrate brain irregularities, such as alien hand syndrome, dissociative identity disorder, and oddness as chronicled by Oliver Sacks in his The Man Who Mistook His Wife for a Hat that just what it means to be conscious and in control of one's body is somewhat malleable.

Journalist Malcolm Gladwell in his Blink: The Power of Thinking Without Thinking built a case of how intuitions sometimes are unknown or even opposed to conscious thought within the same person who is an expert at some forms of knowledge-how. He talked about how high performing athletes sometimes have conscious understandings of their bodies that are the opposite of how their bodies actually function, for instance. Thinking, Fast and Slow by Nobel-prize winning Daniel Kahneman also makes an excellent case for a model he expanded on which categorizes human knowledge into a fast, erroneous sort (Type I), and a deliberate, logical and linguistic sort (Type II).

Ultimately, the relationship between consciousness (SEP) and self-knowledge (SEP) are intertwined, and for the philosopher who embraces science, evolutionary epistemology (SEP) and other modern notions of epistemology promise to develop philosophical theses regarding knowledge along lines of empirical evidence. Philosophers will continue to split hairs over definitions, of course, but they stand to benefit from claims that have scientific certainty about the structure and behavior of the human brain and mind as well as language use. In this way, philosophers can appeal to physical facts to help continue with their taxonomy of knowledge and its sources.


I would argue that an agent, even a non-conscious one, can be described as knowing things. We might say a Roomba knows how to clean the whole carpet. It doesn't know, how to cope with a suddenly opened door, and a world outside without walls.

The word 'consciousness' comes with a lot of baggage, as probably the single biggest source of dispute in modern philosophy. So I would suggest you need to look functionally at what is happening when information becomes knowledge.

I suggest that it is about organising the information so as to be useful, that is getting situated relationally to it. Discussed in more detail here: As humans, do we require a total understanding of information to fully embody it as knowledge? We can use Hofstadter's idea of strange-loops to understand there is something unusual that happens when an agent can include a model of itself in the process of situating - that this can result in deciding how to be (what kind of agent, for desired results), and so support ideas about having intentions (sets of ideas what agent to be, related to goals).

For Hofstadter, a strange-loop investigating self & world in multiple ways & using different approaches, can form a 'tangled hierarchy', which we can understand as the basis of the properties we associate with consciousness - an agent, that is situated in a meaning-cosmology that allows it to act intentionally.


Obviously knowledge is more than data storing: An encyclopedia has no knowledge, but stores data.

I take the term „conscious“ in your question in the sense of „self-conscious“. Then an autonomous Mars rover has knowledge about his environment. Because it navigates on the basis of data it succesively acquires when moving around. But the Mars rover is not self-conscious because it does not consider itself from a metalevel.

I am undecided whether conscious (in the broad sense) processing of data is required for knowledge. I am curious to learn what others think about this point.

Summing up: Self-consciousness is not necessary for knowledge. Knowledge is more than data storing. Knowledge includes the capability to apply the stored data, e.g. for action.

  • What is the delineation for metalevel? If I focus on myself, I have self-awareness. If I focus on others' thoughts, I have theory of mind. If I focus on others' thoughts of me, have I self-consciousness?
    – Michael
    Apr 5, 2022 at 7:20
  • @Michael Assume there are several levels of mental processes: One the basic level 1) I have sense perceptions, emotions, start a motor action etc. . On the higher level 2) I am aware of and discriminate the different acitivities from level 1). On the highest level I am self-conscious and register myself as a unit. Each higher level is a metalevel with respect to the lower levels.
    – Jo Wehler
    Apr 5, 2022 at 8:32
  • +1 Would you accept that an encyclopedia stores knowledge, but has no knowledge of the knowledge it stores, particularly presuming knowledge is the use of information, that is contextualized data, to achieve goals?
    – J D
    Apr 5, 2022 at 14:55
  • @J D I accept that an encylopedia has no knowledge of the data stored in it, because an encyclopedia cannot process those data. But also Searle’s Chinese room has no knowledge of the Chinese language - nevertheless it processes the input data. Hence goal orientation of the processing seems an important requirement for knowledge. The Mars rover acts goal oriented.
    – Jo Wehler
    Apr 5, 2022 at 15:15
  • Technically any input-output function or mechanism has the implicit goal of satisfying its nature. A Chinese room may thus have a goal.
    – Michael
    Apr 5, 2022 at 17:42

Yes, knowledge requires consciousness. Attributing knowledge to a mechanical object like a Roomba is anthropomorphism--the unjustified assignment of human characteristics to things that are not human. The ancient form of anthropomorphism is often mocked today, but the idea that machines think or have knowledge is just the modern manifestation of this ancient error.

If you are conscious, then you have awareness, feelings, thoughts; there is something it is like to be you. No one else can really know what it is like to be you--not fully--but nevertheless there is such a thing. Knowledge is just an aspect of what it is like to be conscious. There is something that it is like to know something. For example, there is something that it is like to know that you are about to die, or about to get laid. It is an internal mental experience, and not merely an outward behavioral pattern.

So, although it is fine in practice to say that a Roomba "knows how to vacuum", you have to understand the statement "know" in that sentence as metaphorical. Because of the way your brain works, when you watch a Roomba, your brain tells you that there is a deliberate purpose there, guided by knowledge, but this is an illusion. The belief that the Roomba really, literally knows something is just as much of a mistake as thinking that a mirage is real; it is being fooled by an illusion.

There is nothing that it is like to be a Roomba. There is nothing that it is like to experience an obstacle like a Roomba. The Roomba operates as it does because of physical forces. It no more knows when it hits a wall than a branch knows when when a bird alights on it, yet the branch droops anyway. Nor does the branch know when the bird flies off, yet it springs up in response. The branch is merely responding to physical forces. The Roomba is nothing more than a very clever device designed to react to physical forces in a way that is useful for vacuuming carpets.


Depends on what we mean by knowledge.

An infant predator (say, a cheetah) isn't born with the instinct to look out for the weaker animals that are easier to pray upon. That's something that it learn from trial and error by watching its mother and role-playing with its siblings.

That's knowledge that is acquired, developed, honed. How much of that is intentional, we do not know, but it is certainly a form of knowledge.

This does not require consciousness (awareness of the self, or awareness that it is aware of said knowledge, or aware that it is aware that it is aware, recursively ad infinitum.)

So, knowledge in this sense does not require consciousness.


The quintessential aspect of knowledge is that it we identify patterns from the data we gather from the world, which help us to simplify our understanding of the world around us. Only sentient beings can gather and process data and extract information from it in order to determine any recognizable pattern in the chaos using their memory and tools. The former implies that only conscious beings can form memory and thus capable of deciphering order from chaos.

In the case of a universe with no beings, its impossible to ascertain, but I tend to gravitate towards "no knowledge in a universe with no beings" view, because knowledge inherently implies a "knower" who acquired information inferred from data and thus who "knows".

  • Machine learning excels at gathering data and identifying patterns, and is not sentient.
    – J D
    Apr 24, 2022 at 13:34

You must log in to answer this question.

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