By LEE J. ELKIN
Essentially, humans have evolved to be able to infer through computing and processing information at a complex level – more than any other biological being. This feature most likely occurred through the process of natural selection according to the theory of evolution, and thus human beings have adapted to such feature. Although it took sometime to develop computational skills, it is proven that humans have adapted adequately tracing back to antiquity based on our evidence provided by historical and anthropological records. And humans have continued to fulfill their cognitive capacities by developing more complex cognitive skills up through the present day. To illustrate this point, let us assume that the first evolved human may not have understood the inference used in the proposition ‘2 + 2 = 4’, but today, a young child could exercise her computational skills and combine two objects with another two objects and come to the conclusion that two and two make up four, though the child might not know the language, depending on her age.
Moreover, it appears that the computational adaptation has become a genetic trait among humans. Since humans are constantly thinking, and there are many out there that devote their time to inquiring especially into scientific problems, which entails rigorous reasoning, inference is a common phenomenon on a daily basis. We use inference in mathematics, physics, logic, chemistry and most, if not all, of the “professional” disciplines, but it is not just limited to these subjects, it is also used in more simple things such as directions, cooking a meal, and other ordinary activities that people partake daily. Sure, what I have given thus far is just a generic description of the use of inference, but what is an inference exactly other than reasoning between things to derive a conclusion? By this, I mean, what is the nature of inference?
The activity of inference concerning one’s experience is usually founded with symbols or objects, and because of external factors playing a fundamental role in the inference, it seems plausible that the phenomenon, in part, has somewhat of a metaphysical nature from a mind-world perspective. My aim in this essay is to explore the nature of inference and attempt to show how it may be possible that inferences made about the external world, in part, are not just something of a physical nature that is contained and occurs only within the brain, which I think most people nowadays believe that it is only a physical process and consciousness all together is contained within the brain. In doing such task, I will inevitably show that inference requires the two branches of consciousness – access and phenomenal consciousness where the latter will draw on the active externalism thesis.
§1. An Explanation of the Nature of Inference (Roughly Speaking)
As I have just stated, the activity of inference based on an experience is usually founded with symbols or objects to begin with. To use a basic inference in logic for example, imagine using modus ponens. I have a conditional statement, then an assertion (the antecedent of the conditional statement) and from that I can derive a conclusion (the consequent of the conditional statement) inferentially.
A → B
Or I could use a simple mathematical expression similar to the one in the introduction – a combination of one and two leading to the result of three:
1 + 2 = 3
In my above examples, I ought to conclude that both examples are familiar cases of simple inference that philosophers and mathematicians are well acquainted with. Each are represented by symbols, but inferences are not limited to just being represented by symbols, they can be represented with objects (through experience) as well. Now, a neuroscientist may advocate that these inferences are a purely physical process occurring within the brain. It might be due to some chemical signal, firing of neurons, and so on. A computer scientist and artificial intelligence supporter would most likely say something similar but in terms of computer hardware. With all of the new discoveries in neuroscience and intelligent systems, it would be hard for one to doubt that inference is entirely of a physical nature that only occurs within the brain (or computer system). However, the problem at hand is the lack of clarity of this being a physical phenomenon entirely. The water is still cloudy so to speak and I cannot be certain that this conclusion is accurate without a more thorough look.
I am not going to deny that the biological components of the brain, in part, play a crucial role in human inference. In fact, because of biological neural networks, information can be transmitted and processed through chemical signals. But there seems to be another part that is missing and is not explainable through these neural systems. The feature, partially being developed by the brain, is something beyond matter when we look at inference from a mind-world perspective. There is a relation between mind and world where certain aspects of the world are an external extension of the mind – meaning that consciousness is not just contained within the head – and this view is commonly referred to as active externalism (1) (Chalmers & Clark, 1998). Similarly, Alva Noë appears to be a supporter of this line of thought as well, but has applied it to experience, and this application to experience is what I will be most concerned with throughout the paper. In “Experience Without the Head,” Noë states:
…what we experience visually (for example) may outstrip what we actually see. From this it follows not that experience could not be in the head. What follows, rather, is that it might not be, or rather, that some aspects of some experiences might not always be. A modest conclusion, but one that allows that, at least sometimes, the world itself may drive and so constitute perceptual experience. The world can enter into perceptual experience the way a partner joins us in a dance, or – to change the image slightly – the way the music itself guides us (Noë, 2006).
Noë makes the modest conclusion of “well, it might be or might not be,” which seems to be “iffy” on the subject, but I don’t think that such modesty is needed. There are instances given that serve as proof for the active externalism thesis (2). And active externalism also appears to be obvious in the case of inferences made about the external world because one is unable to make an inference on the most basic (or complex) events that occur within the world if there are no external objects (or symbols of representation present) given to make an inference from. The objects themselves give us a starting point to make inferences.
One may argue, however, that we do not need external objects (nor symbols) given to us through experience to make an inference relevant to the external world, but rather, some inferences can be made based on innate knowledge (3), or the protester may attempt to coin this knowledge as a priori, but it is obvious that ‘a priori’ here is being used in the wrong sense. Moreover, it seems absurd to believe that we have any sort of innate knowledge, especially with what empiricism has taught us over the past few centuries. It does not seem absurd, however, to have a priori synthetic knowledge in the correct sense. What I have expressed as a counter-argument above does not address a priori knowledge in the correct Kantian sense. If we were to take Kant’s actual proposal of there being a priori synthetic knowledge, then yes, I do believe that an inference can be made independent of perceptual experience, but this is much more difficult to explain, which would entail an extensive analysis, and is a different topic for a different day. We can conclude, however, that Kant would agree that experience “awakens” our faculty of knowledge and, for our sake, a priori synthetic knowledge will have to take the backseat to experience as far as we will be concerned. What I am trying to get at is explaining how we make inferences about the external world in a general sense and not to nit pick at our knowledge of space, time, geometry and the like – I am taking for granted that these types of knowledge are already presupposed. Therefore, experience must play a role in the foundation of inferences made about the external world. Once there are objects that are given to one’s perception for that person to make an inference on, the inference, in part, appears to be a quale. Why does it appear to be a quale? It is because there is an unexplainable, sensational gap, in terms of matter, between mind and world at this point in time. Let me use a diagram to illustrate this:
Now the use of the term ‘gap’ (also referred to as the “explanatory gap” by some) is ambiguous here, but what I think the gap, preventing consciousness being declared one-hundred percent fully physical due to the brain, is qualia and in this case, inference, in part, counting as a quale. The problem here is a sensational one, and inference can be partially of access consciousness, but I believe that it can also be partially of phenomenal consciousness (4) (Clark, 2000), where the phenomenal feature of inference is the unexplainable part in terms of matter. The reason for it being split into two parts is due to the one aspect being part of the computational (access) – cognitive processing – and the other being sensational (phenomenal) through perceptually experiencing objects, which contain different qualities that can be sensed, within the world to make inferences on.
When one perceives say a man and woman kissing, she would infer that they are a couple (5). The image would be processed within the brain to make the inference that they are a couple, but the inference begins before the brain receives the image because the experience of the event is necessary to make the inference at all (the perceptual experience is the foundation of the inference) and this is where the unexplainable gap of inference occurs sensationally, or if it were not the case that certain aspects of the world function as an extension of the conscious mind by aiding our mental processes, then this inference would not be possible to make because one would not have the representational mental image of the man and woman kissing. Thus, the cause of the inference does not begin with computation in the brain; the cause of inference in most cases lies in one’s perceptual experience of the external world.
With this explanation of the two parts of consciousness being required in making an inference relevant to the external world, it appears that I am taking a Kantian approach. As Kant had described in his Critique of Pure Reason, the faculty of knowledge is distinguished by sensibility and the understanding. Objects are given to us by means of sensibility, which yield us intuitions, and the understanding allows intuitions to be thoughtful, which concepts arise. The effect of an object on the faculty of representation is called sensation. And the intuition that is in relation to an object through sensation is empirical. So, an empirical intuition gives us appearances of objects that can later become knowledge (Kant, 1929). If we stand on this latter notion only, however, we might run into a problem with giveness or simply putting it, fall into Sellars’s “Myth of the Given.” Luckily for us, that is not the case. Even though empirical intuition gives us appearances of objects, it cannot by itself constitute knowledge. Rather appearances are thought spontaneously through the understanding and concepts of the appearances develop from the understanding. Thus, I think it is safe to say that through concepts, we can have knowledge. Similarly, my approach to inference takes a resembling line of thought with emphasis on a holism of phenomenal and access consciousness being relevant in one making an inference. By only having the perceptions of objects, we cannot call that knowledge. Not only is that problematic, but the perceptions cannot be computed through phenomenal consciousness and thus we have no inference at all. In support of this idea, imagine perceiving an object at one instance and the object later becomes a concept that is cognitively accessible. If the object changes over time and we perceive it after the change, then we have a new perception of the object, which becomes a concept and clashes with the previous concept of that object. So the latter would replace the former, I think, and we would have a new cognitively accessible concept. However, such a process cannot occur in phenomenal consciousness because we would not recognize the change in the object that occurs between perceptual instances. So there is a dependence on access consciousness needed here to process images and output a conclusion inferentially. The dependence, however, is necessarily reciprocal, and that is the stressing point of this paper.
There still may be confusion lingering around on how inference, in part, can be a quale since what I have mentioned thus far are objects in general and not subjective qualities. However, I will attempt to address this issue now. When one perceives anything in the world, the perceiver’s experience of objects leads her to perceive qualities of the objects as well. The qualities are contained within all objects and necessarily predicate those objects. I do not think that it would even be conceivable to imagine an object absent of “secondary” qualities. For example, if a mysterious object had appeared to me in which the object looked though as the color was absent due to some aspect of the object being unfamiliar to me, I would be wrong in allowing myself to think this way because my perception surely could be matched up somewhere on the color spectrum. Moreover, since objects are the foundation of inferences made about the external world, and all objects necessarily contain various qualities within themselves, then that implies that the cause of inference (perceptual experience) necessarily contains qualia, and thus qualia is a necessity in phenomenal consciousness of an inference, which would make itself a broader version of qualia accounting for every sensational property and not just an individual sensation such as ‘seeing red’.
§2. An Attempt of Clarification
Clearly, my theory would seem obscure to most and it could be due to misunderstanding the theory since it is quite a task to explain it coherently6, but let me attempt to explain it analogously to a popular thought experiment. In Frank Jackson’s thought experiment about Mary and the black and white room, Mary was confined to a black and white room and was never exposed to any other colors. She knew everything there was to know about the physical world through her books on chemistry, physics and neurophysiology. One day, Mary was taken out of the room and shown a ripe, red tomato. Mary had no idea what the property red was since she had never experienced it before. Therefore, Mary learned something new and proved that she did not know all there is to know about the world (Jackson, 1986).
Similarly, imagine mechanical Mary, who is exactly the same as a human, in a room that is completely dark with no lighting of any sort. A constant temperature is kept so that Mary cannot sense a variation. There are no sounds, smells or things to be tasted in the room and the scientists of the experiment have developed devices that would temporarily disable Mary’s sense of taste, smell, touch and hearing, but she would remain a conscious being. Since the room is extremely dark, Mary would not be able to use her sight to view anything other than blackness. After several years, mechanical Mary is taken outside of the room and exposed to the real world and the scientists re-enable her senses. Obviously, she has no clue of what anything is that she perceives. But since she is like a human in every way, she has computational skills programmed in her. The scientists sit her at a little children’s play table with two wooden blocks on one side and two on the other. Mary sees the one set of blocks and instinctively moves her arm towards the blocks. The back of her hand smashes into the blocks causing the blocks to fall on the floor. Like a child, Mary recollects on the act and impulsively lets out a laugh, and then she begins to move her other arm towards the set of blocks that remain on the table and proceeds to push them onto the floor as well because she thought that the act was funny and it made her feel happy. So she inferred that by doing it again, she would obtain the same result.
From this experiment, the scientists have learned that mechanical Mary performed a basic computational inference through performing an instinctive act, which caused a mess and made Mary laugh and feel happiness for the first time. She had inferred by doing the same procedure again to the other set of blocks, she would obtain the same result. However, while confined to the black room, she had never inferred anything of this nature before other than the fact that she exists, but nothing about the actual world and her environment because she had never experienced anything sensationally beyond staring at the blackness of the room. Thus, an inference about the external world is dependent on perceptual experience. The world provides objects that eventually become representational mental states, which are sent and computed in the brain for one to make inferences on. Let us not forget that the objects give rise to non-representational mental states also. And since this type of inference is dependent on objects, inference extends over mind and world, which would entail inference partially due to cognitive processes and partially due to sensation where the latter is unexplainable in terms of the physical because of subjectivity and qualia being contained within the experience, which are necessary components in the perception of objects, and I think inference can be declared as a quale itself in certain respect on the phenomenal consciousness side.
Furthermore, the thought experiment regarding mechanical Mary also illustrates that the success of Mary’s adaptation to her environment is dependent on the active externalism thesis. Even though this is a fictional case, we are shown that if it were possible to insert a human into the situation with all of the necessary conditions, then we could conclude that the success of the adaptation to the environment is dependent on the environment providing the human with external “instruments” that would serve as an extension of the person’s consciousness by aiding their mental processes, which will allow them to make inferences.
To conclude, I have argued that inferences made on worldly experience rely on a holistic process. Not only is access consciousness, which does the work of computing information to output a conclusion, but phenomenal consciousness, where the inference begins, is required also since it is the starting point of the inference itself. Overall, I believe that this new analysis of inference has created another roadblock for physicalism since phenomenal consciousness entails subjectivity and qualia in perception. Qualia have been problematic for physicalists in the philosophy of mind and also have prevented a plausible materialist position of the mind from being developed.
Chalmers, D. & Clark, A. (1998) “The Extended Mind,” Analysis 58, pp. 10-23
Clark, A. (2000) “A Case Where Access Implies Qualia,” Analysis 60: 265, pp. 30-38
Jackson, F. (1986) “What Mary Didn’t Know,” Journal of Philosophy 83: 5, pp. 291-295
Kant, I. (1929) Critique of Pure Reason, Translated by Norman Kemp Smith, New York, NY:
Palgrave Macmillian, pp. 65-67
______(1977) Prolegomena to Any Future Metaphysics, Translated by James Ellington,
Indianapolis, IN: Hackett, pp. 10-11
Noë, A. (2006) “Experience Without the Head.” In Tamar Szabo Gendler & John Hawthorne
(Eds.) Perceptual Experience, New York, NY: Oxford University Pres
Sellars, W. (1997) Empricism and the Philosophy of Mind, Cambridge, MA: Harvard University Press
(1) This is a generalization made of Chalmers and Clark’s theory of the extended mind.
(2) See Chalmers & Clark, 1998. The Tetris case and the thought experiment regarding Otto who has Alzheimer’s.
(3) One might question why I have brought up innate knowledge. They may assume that I am still dwelling on the Cartesian notion of the mind having some innate knowledge, but rest assured that is not the case. It is the case, however, that some linguists, psychologists, cognitive scientists, and even philosophers today still believe that the mind has some innate concepts. Thus, this counter-argument is not all that improbable.
(4) See Clark, 2000. Access consciousness, in brief, is when content of a mental state is available for control of rational action for use of verbal reports or reasoning. In contrast, phenomenal consciousness involves experiential properties such as ‘what it is like’ to see red and the like (qualia). This is a thesis that has been defended by Ned Block and others.
(5) The inference need not be true. P & Q will φ iff P & Q are a couple. It is not the case that this statement is necessarily true. P could be a man where Q is his mother, which would entail, in a strictly normal sense, that they are not a couple. In the latter case, the above statement would be false. Moreover, I am not particularly concerned with the truth or falsehood of the inference, just the inference in general.
(6) As you will notice, section one is somewhat scattered and ill-structured. But when read in its entirety, I think it forms a bigger picture.
L.J. Elkin (’09) is a Philosophy major at University of Pittsburgh.