How Similar are Imagination, Memory, and Perception?

 Bartek Chomanski is a Visiting Lecturer in Philosophy at Northeastern University. His interests span a range of topics in philosophy of mind, with particular emphasis on connections between consciousness and spatial representation.

Bartek Chomanski is a Visiting Lecturer in Philosophy at Northeastern University. His interests span a range of topics in philosophy of mind, with particular emphasis on connections between consciousness and spatial representation.

A post by Bartek Chomanski.

In this post, I argue that the phenomenological similarity between imaginative, recollective, and perceptual experiences is a complex matter, and that accounting for it is made difficult by empirical data which suggests that there is closer similarity between imagination and perception, and imagination and memory, than there is between memory and perception. I then explain why I think the empirical results raise a potential problem for a certain style of accounting for this similarity and sketch some solutions.

To avoid complication, I will talk about instances of sensory, episodic visual memory, visual imagination, and visual perception. I leave these notions unanalyzed, as I think an intuitive understanding of them will suffice.

With these preliminaries out of the way, I will now develop my argument.

First, consider the following claim that I’ll call the Broad Similarity Thesis (BST):

BST: Recollective, imaginative, and perceptual experiences are all similar to one another in phenomenologically salient respects.

BST is, I think, among some of the basic introspective data we have about the phenomenology of the three types of experiences under discussion. Thus, one part of the challenge for a philosophical account of the phenomenology of imagination (or memory, or perception) is to explain how the three types of experience differ, and how they resemble each other (see for example Byrne, 2010; Kind, 2017).

Empirical research, as well as introspective reflection, suggest that the similarities and differences between imaginative, recollective, and perceptual experiences are not perfectly symmetric. In particular: confusions between perceptual and imaginative experiences (“P/I confusions”) are well-known and well-reported (Perky, 1910; Segal & Nathan, 1964). One can also find indications in the literature that confusions between recollective and imaginative experiences (“R/I confusions”) do happen (Rubin et al., 2003). Yet, it is difficult to find a case in which a subject confuses a perceptual experience of an object with a recollective experience of that object (“R/P confusions”). This is prima facie evidence of second-order dissimilarity: the similarity between P/I is more like the similarity between R/I than it is like the similarity between R/P[1].

Now, we shouldn’t immediately think that the absence of R/P confusions is a reason to reject BST. Similarity between Fs and Gs needn’t imply that it’s possible to confuse an F for a G. Moreover, the possibility of confusion between an F and a G, and between a G and an H, should not imply that it must be possible to confuse an F and an H (after all, there could be confusions between yellow and orange, and orange and red, without there being confusions between yellow and red).

That said, I will assume that the actual instances of R/I and P/I confusions suggest that there could be borderline cases between a recollective and an imaginative experience, as well as between a perceptual and imaginative experience (“borderline R/I cases” and “borderline P/I cases”). I will also assume that there are no borderline cases between recollective and perceptual experiences (“borderline R/P cases”).

The existence of borderline R/I and P/I cases, but no borderline R/P cases, suggests a constraint on those accounts of imaginative/perceptual/recollective phenomenology that want to adhere to BST. It suggests that such theories should not allow for the possibility of borderline P/R cases, and that they should allow for the possibility of borderline R/I and P/I cases.

For example, consider these three experiences: a sensory memory, a visualization, and a perception of two birds fighting. It is tempting to explain the phenomenological similarity between these three experiences by saying that they all share some core content (that the birds are fighting). It is tempting to explain the phenomenological differences between these experiences by saying that each type of experience involves a distinct additional element (such as a sense of pastness, or that the situation is possible, or that the perception is caused by the objects it is a perception of; or that each of the three types of mental state involves a specific way of entertaining the core content, etc.). Let’s call such views Content+ (C+) views.

C+ views must allow, to be consistent both with BST and the empirical findings I mention, that whatever aspect[2] of an experience it is that distinguishes imagination from memory can be so manipulated that a borderline case arises; and whatever aspect of an experience it is that distinguishes imagination from perception can be so manipulated that a borderline case arises. If this is right, then simultaneous manipulation of both aspects at the same time is possible, and, if so, it should make possible a borderline case between a perception and a memory. But such cases don’t seem to happen. This, I think, is a problem for C+ views.

In standard form, my argument is this:

(1)   Suppose that similar recollective, imaginative, and perceptual (R/I/P) experiences share the same content. (part 1 of C+, to explain BST)

(2)   Suppose that the R/I/P experiences differ by each having some additional aspect(s) that the remaining kinds of experience may lack. (part 2 of C+, to explain phenomenological differences)

(3)   If (1) and (2) are true, then it is possible to transition from any one experience-type to another by addition/removal of the right kinds of aspects to the content of the experience.

(4)   Either the addition/removal is gradual or instantaneous.

(5)   If it’s possible to transition from any one experience to another by gradual addition/removal of the additional aspects to the content of the experience, then there should probably be borderline cases between all R/I/P experiences.

(6)   But there are no borderline P/R cases.

(7)   If the addition/removal is instantaneous, then then there should probably be no borderline cases between all R/I/P experiences.

(8)   But there are borderline R/I and P/I experiences.

(9)   Therefore, probably, at least one of (1) and (2) is false.

Many replies to this are possible. To give a few examples:

(a)   Perhaps what explains BST and the data is that imagination is actively involved both in memory and perception (see Nanay, 2010 for the latter), but memory is not involved in perception.

(b)   Perhaps it’s illegitimate to assume there are borderline R/I and P/I cases.

(c)   Perhaps there’s something special about additional elements of each type of the experience that makes gradual R/P transitions unlikely, while allowing for gradual P/I and R/I transitions.

(d)   Perhaps Perky-style experiments test imagery or visual memory – and so, arguably, don’t pose a threat to BST – rather than visual imagination.

(e)   Perhaps episodic memory is merely a kind of visual imagining (see Langland-Hassan, 2015).

These could be plausible lines of defense against (or accommodation of) my argument and working out their details could advance our understanding of the phenomena involved.


[1] This isn’t the only asymmetry. Another is that it’s possible to imaginatively interject imagined objects into a perceived scene, and into a remembered scene, easily and at will. However, it’s much more difficult to interject a remembered episode into a perceived scene (unless one perceives the scene where the remembered episode took place).

[2] The C+ views need not be committed to there being only a single aspect that distinguishes the relevant experiences. I adopt this manner of speaking merely for ease of formulation.


References:

Byrne, A. 2010. Recollection, perception, imagination. Philosophical Studies148(1): 15-26.

Kind, A. 2017. Imaginative Vividness. Journal of the American Philosophical Association, 3(1):32-50.

Langland-Hassan, P. 2015. Imaginative attitudes. Philosophy and Phenomenological Research90(3): 664-686.

Nanay, B. 2010. Perception and Imagination: Amodal Perception as Mental Imagery. Philosophical Studies, 150(2): 239–254.

Perky, C.W. 1910. An experimental study of imagination. The American Journal of Psychology21(3): 422-452.

Rubin, D.C., Schrauf, R.W. and Greenberg, D.L. 2003. Belief and recollection of autobiographical memories. Memory & cognition31(6): 887-901.

Segal, S.J. and Nathan, S. 1964. The Perky effect: Incorporation of an external stimulus into an imagery experience under placebo and control conditions. Perceptual and Motor Skills18(2): 385-395.