Cognitive Phenomenology: The Stream of Consciousness

According to William James experiences, including conscious thoughts, flow in a stream of consciousness. Peter Geach argued that whatever we say about other experiences, conscious thoughts at least do not flow, but rather occur in discrete sequences. A number of recent arguments against cognitive phenomenology take Geach’s criticisms of James as their point of departure. In the book I examine them in detail and give reasons to think none works. Here I will discuss the background to these arguments, showing that Geach himself at least provided no good reason for skepticism about cognitive phenomenology.

What is a Jamesian stream of consciousness? Let’s say that a minimal stream of consciousness is a binary relation between times and experiences. A Jamesian stream of consciousness is a minimal stream of consciousness that also meets the following constraints:

Synchronic Unity: The experiences a stream of consciousness assigns to one time tend to be phenomenally unified.

No Constancy: If t1 and t2 are different times in the domain of a stream of consciousness, then there is some experience the stream assigns to one of t1 or t2 but not the other.

Diachronic Unity: For any time t within the domain of a stream of consciousness, if there is any interval around t, then there is an interval I around t such that the experiences the stream assigns to t tend to be phenomenally unified with the experiences the stream assigns to other times within I.

These come from an analysis of James’ discussion in The Principles of Psychology. I won’t try to defend the analysis here, but just assume it is accurate.

Now let’s look at what Geach says. In his paper “What Do We Think With?” (reprinted in God and the Soul) he writes:

Thinking consists in having a series of thoughts which can be counted off discretely—the first, the second, the third, …—; which, if complex, must occur with all their elements present simultaneously; which do not pass into one another by gradual transition.

Geach explicitly contrasts this view with James’s: “thoughts occur not in a Jamesian stream, but as I maintain that they do—as a series in which certain thought-contents successively occur; with no succession within any one thought and no gradual transition from one thought to another…”

The challenge to cognitive phenomenology might be put like this:

(1) If thoughts have irreducibly cognitive phenomenal characters, then they must occur in a Jamesian stream of consciousness

(2) But as Geach points out, thoughts do not occur in a Jamsian stream of consciousness.

(3) So thoughts do not have irreducibly cognitive phenomenal characters.

Let’s grant (1). Premise (2) is mistaken. Geach’s plausible observations about the sequential character of thinking do not imply that thoughts cannot occur in a Jamesian stream of consciousness. This is obvious once you make Geach’s observations more precise.

Geach makes two observations about the sequential character of thinking. Here are sharper formulations:

Simultaneity: Suppose you think that p at t. Then all the parts of your thought that p occur at t.

Disjointness: Suppose you think that p at ta and think that q at a distinct time tb. Then there is no overlap between your thought that p and your thought that q—i.e. your thought that p and your thought that q do not share any parts.

Simultaneity is the claim that all the parts of a thought occur simultaneously and Disjointness is the claim that there are no gradual transitions between thoughts.

Neither of these claims is inconsistent with the claim that thoughts flow in a Jamesian stream, as I have characterized such a stream. To make the separation of these issues clearer, let us consider how Geach’s view might be altered so as to generate claims that are clearly inconsistent with the claim that thoughts flow in a Jamesian stream.

I will focus on Disjointness, as it is more obviously related to this issue. Here is a first pass:

Disunity: Suppose you think that p at ta and think that q at a distinct time tb. There is no relation of diachronic unity that obtains between your thought that p and your thought that q.

Jamesian streams flow in virtue of relations of diachronic unity, not in virtue of relations of mereological overlap. So to generate an inconsistency we want something like Disunity. But Disunity is not enough. The reason is that James’s commitment to Diachronic Unity is about all sorts of items in a stream of consciousness, but Disunity is about thoughts. Consider the way that the natural numbers occur in the real numbers. They are a discrete subsequence within the richer continuum. Similarly, thoughts might be a discrete subsequence within the richer continuum of experiences in general. If Disunity were true, then there could be no Jamesian stream of consciousness consisting wholly of thoughts, but thoughts could occur within richer streams of consciousness that have their stream-like, flowing character in virtue of diachronic unity relations between items in the stream in general.

It is not clear what motivation there is for holding this view. What bars thoughts from standing in diachronic unity relations to each other? One might argue that diachronic unity relations obtain in virtue of mereological overlap relations. Some people defend this view, but it seems implausible to me. A visual experience can be diachronically unified with a gustatory experience, even though they share no parts. Explaining phenomenal unity in mereological terms is an attractive idea, but it is possible to do so without requiring unified experiences to overlap. One idea is that two experiences are unified if they are both parts of a larger experience. They do not need to share parts. Rather, they need to be parts of a larger experience. This sort of view is compatible with diachronic unity relations holding between non-overlapping thoughts. So Disunity is rather questionable. Still, it is an interesting thesis, worth flagging. It implies that even if thoughts do occur in a stream of consciousness, they could not constitute a stream of consciousness on their own.

To rule out the possibility of thoughts occurring in a Jamesian stream of consciousness at all, we would have to strengthen Disunity to the following:

Discontinuity: Suppose you think that p at ta and have an experience, e, at a distinct time tb. There is no relation of diachronic unity that occurs between your thought that p and your experience, e.

If Discontinuity is true, then thoughts introduce disruptions in a stream of consciousness, no matter what other sorts of items that stream contains. This is inconsistent with James’s view that the stream of consciousness obeys Diachronic Unity. It is unclear what the motivation for Discontinuity might be.

Let me emphasize that I am not rejecting either of Geach’s claims—Simultaneity or Disjointness. Both seem plausible to me. What I am denying is that their truth immediately renders thoughts unsuitable to be parts of a Jamesian stream of consciousness, where such a stream is characterized by Synchronic Unity, No Constancy, and Diachronic Unity.


  1. Do you really think the term “thought” has a definite enough meaning to support arguments of this sort? As far as I can see, a “thought” is more or less an unspoken sentence or sentence fragment. Do you have something more definite in mind? (Perhaps this is dealt with in your book or Geach’s?)

    To illustrate the problem, let me add a passage from Joyce’s Ulysses that I have several times used to illustrate the concept of stream of consciousness:

    Yes because he never did a thing like that before as ask to get his breakfast in bed with a couple of eggs since the City Arms hotel when he used to be pretending to be laid up with a sick voice doing his highness to make himself interesting for that old faggot Mrs Riordan that he thought he had a great leg of and she never left us a farthing all for masses for herself and her soul greatest miser ever was actually afraid to lay out 4d for her methylated spirit telling me all her ailments she had too much old chat in her about politics and earthquakes and the end of the world let us have a bit of fun first God help the world if all the women were her sort down on bathingsuits and lownecks of course nobody wanted her to wear them I suppose she was pious because no man would look at her twice I hope Ill never be like her a wonder she didnt want us to cover our faces but she was a welleducated woman certainly and her gabby talk about Mr Riordan here and Mr Riordan there I suppose he was glad to get shut of her

    Does that constitute a discrete or continuous stream of thought?

    Best regards, Bill

    • Eli

      Hi Bill,

      You are exactly right that one needs to spend some time circumscribing the mental states to which Geach’s observations apply. I do go into this in the book. I take Geach’s observations to apply to mental states that just have a propositional attitude structure: there is the attitude–e.g. believing, desiring, entertaining–and the propositional content–e.g. that he hasn’t asked for eggs with his breakfast in bed since staying at the City Arms hotel–and that is it. Geach’s idea is that if this is the constitution of a mental state, then it is a mental state one is in just in case one bears the right attitude to the right content and that is an all or nothing affair. Because there are cognitive states that have more structure than this, even if Geach’s claims about the stream of consciousness held up, the result for cognitive phenomenology would be limited.

      About the Joyce passage: the interior monologue might be continuous but still associated with a discrete sequence of propositional attitudes. It might be difficult to figure out exactly which propositional attitudes from Joyce’s words–and, well, since this is a fiction, the question probably does not have a fully determinate answer.

  2. The concept of a mental state seems just as fuzzy to me as the concept of a thought, so let me try to frame a more concrete question.

    (Forgive me if I’m telling you things you already know: ) The Canadian psychologist Donald Hebb wrote a very well-known book, published in 1949, called The Organization of Behavior . He proposed that the structure of thought can be understood at the brain level in terms of two concepts that he called cell assemblies and phase sequences A “cell assembly” is a group of brain cells that once activated can remain active for some period of time because of mutually reinforcing connections between them. A “phase sequence” is a sequence of cell assemblies, each of which is active for some time and then transitions discretely to the next one.

    We don’t understand the human brain well enough to know whether Hebb was correct, but my question is whether, if he was correct that cognition at the brain level consists of a sequence of discrete states, it would mean that Geach is correct and you are wrong. Or is the issue between you and Geach framed at the different level?

  3. Eli

    Hi Bill,

    I agree with Geach that thoughts occur in discrete sequences. I disagree with his claim that this renders them unsuitable to be occupants in a Jamesian stream of consciousness.

    Aside from that, however, I think there can be a mismatch between what we say about brain processes and mental states. Suppose brain processes are continuous. Still they might constitute a discrete sequence of mental states, maybe one counted off each time a threshold of some sort was met. Suppose instead that brain processes are discrete. Still they might constitute a continuously enduring mental state, maybe a mental state one counts as being in every moment during which some discrete brain processes is clicking along.

Comments are closed.