Theory of universals

Metaphysics and Epistemology
1

I have a trouble understanding Objectivist theory of universals. The essence of this trouble is that I don’t see how it doesn’t collapse into nominalism and then subjectivism.

As I understand, the core difference which Objectivists claim to have with nominalists is that they hold concepts to objective, i. e. fully grounded in reality only. Objectivists even suggest the concrete procedure by which concepts can be derived from reality: measurement omission. It is supposed that we differentiate a set of entities (or relations, or actions) which have a common attribute and then integrate this set in a concept by omitting the concrete quantity of this attribute. As attributes are objective, concepts, discovered by this procedure, are objective too.

The problem which I have with this approach is that I didn’t find a clear definition of “attribute” anywhere in the Objectivist literture I have read. Therefore, I don’t see any intellectual procedure by which it is possible to distinguish clearly between two distinct attributes and different quantities of one attribute. Colors are the best example here - when we stop to see the different quantity of red color but start to see the orange one? And another question which is even more tough: when we form the concepts of particular colors, “redness”, “greenness”, “yellowness” are considered different attributes. But when we form the concept of “color”, they suddenly appear as different quantities of a single attribute - color. And how then attributes are metaphysical, not epistemological as miss Rand emphasized (see ITOE, Appendix)?

Then further difficulties arise. Not all concepts are formed by measurement omission. Miss Rand herself stated that while discussing the process by which the concept of attribute is formed (this is somewhere in the Appendix to ITOE). And in some cases measurement is not omitted entirely. For example (I don’t have a book nearby for reference, so I may be wrong here) while explaining how we form the concept of “table” she admitted that we include in this concept objects of limited size only (for example, flat building raised to the sky on four huge pillars will not qualify as table). The criterium by which we can distinguish between correct and incorrect concepts seems to be lost here.

These problems are partially addressed by Leonard Peikoff when he discusses objectivity of concepts in OPAR. As I understand him, he points to “cognitive need” as the ultimate criterion by which we should judge the correctness of the concepts. I have several problems with this answer.

First, I don’t see, how it is different from “convenience” for which Objectivists mock nominalists (and after all qualifications I also stopped to see the difference between “measurement omission” and “family resemblance”). Second, it seems to imply circularity: for then we must know the nature of things in a set to know if it is necessary to integrate them in a concept, but how can we know their natures before we conceptualized them? It also seemed to me the week point of pragmatism because evaluation should be preceded by conceptualization. I first thought that Objectivism solved this problem by introducing measurement omission theory where concepts depend on directly observable qualities of things only, but now it looks like that Objectivism commits the same error.

I think that’s all for now. The text is rather disorganized, sorry if it creates difficulties. I can feel that all problems which I listed have some single premise behind (wrong premise, I really hope, for I don’t want to prove Objectivism wrong), but I can’t name it, hence the poor structure. As for now, it is one of the two points, where (in my judgment) I don’t understand Objectivism (hope to start a thread on another point in a few days). I had much more in the past and every time it turned out that the reason behind my lack of understanding was unknown acceptance of some premise of traditional philosophy. May be this is the case this time too.

I’m not a native speaker and it would be hard for me to express these thoughts even in my own language. So please don’t judge my poor English harshly.

2

Objectivism (correctly) considers the “problem of universals” to be a pseudo-problem. The main problem of “universals” lies in the dependence on a quasi-Platonic notion of reified attribute as primitive building block. To take your color example, “red” and “orange” are grounded in sensory experiences that arise automatically, given external objects and the visual organs. Whether or not the various units are integrated or differentiated depends on a prior decision to conceptualize – sometimes they are unified into a single concept. Whether or not there is differentiation depends on the cognitive benefit of conventionalizing a difference. Since I think the concept “color” is actually formed by unifying existing color concepts like “red”, “green”, “white” etc. (as opposed to red vs. orange which is a process of further differentiation), this concept would fall under the rubric “higher abstraction”. Although people are not naively aware of the wavelength basis of color, there is certainly a perceptible similarity between the various colors which would preclude forming a concept from various colors plus the sound C sharp.

I don’t see that marginal cases pose any problem for concept formation via measurement omission. It may be important to understand, though, that “measurement omission” does not mean “without regard for any measurable quality”. For instance, “red” refers to a range of wavelengths, meaning “any in that range”, but nevertheless it must be within that range. The broad characterization of “table” as “flat surface on a support” is not complete, hence the building example fails on grounds of size and purpose.

As for the optionality of concept formation, recall also that knowledge is incremental and also not instantaneous. For instance, I imagine most people here do not have any need to distinguish “all-white reindeer” from “third-year male reindeer” or “second-year male reindeer” at the conceptual level, thus one can use compositional expressions to talk about such things, if you ever had to. Some people need to make those distinctions on a regular basis, hence they are expressed economically as words (gabba, vuobis, varit). The ability to detect the differences is there universally; the need to formalize that relation depends on the objective nature of the things and the people doing the conceptualization.

3

One of the problems I see arising consistently with those new to Objectivism is the lack of understanding of the role of observation in concept formation. We perceive reality directly, and concepts are formed from the observation of differences and similarities. The process is objective insofar as it is actually based on the identification of the difference between an aspect of existence that is contextually different from the rest of the awareness field and then integrated together using measurement omission.

For example, let’s say we have the following string of symbols: . - A A , /

Obviously, on the perceptual level, the two “A” symbols are similar, as they only differ by position. So one could form a concept uniting the two A’s together, and calling the concept for that similarity the letter A.

Later, one notices that, compared to the smooth texture of the background color of the screen where one is seeing these, that they all stand out and one can form the wider abstraction symbol, or alpha-numeric symbol, which would include all the character written in this post, each considered individually.

Just hitting keys randomly on my keyboard produces: q f o r 6 7 & * w : " ’ t

As differentiated from the background, these are all similar as they are all black shapes on a white to off-white background.

So, the objectivity comes in on two different levels, the level of perception, in this case an obvious distinction between this –><– and something in the considered range [e w t f] that stands out from the background; and at the level of doing something with that observation with the mind, such as giving them a concept to distinguish them from nothing on the paper, so to speak.

In other words, on the lowest level of concept formation, differences and similarities are given in perception, and one does not need to use any cognitive skills to see them. The sense are automatic, and clearly give a difference between a space versus a symbol.

It is when one gets more and more abstract that one has to use cognitive abilities because one needs to keep a lot in mind that isn’t readily available via perception; i.e. the difference between electrons and protons are not given in perception, but require a whole conceptual context in order to be understood as protons being different from electrons, but all electrons are similar to other electrons and all protons are similar to all other protons. And there is a wider similarity between electrons and protons in that they are each charged particles, versus say, the neutron, which has no charge.

4

David, thanks for answers!

The main problem of “universals” lies in the dependence on a quasi-Platonic notion of reified attribute as primitive building block.

True. But then, what is attribute? What’s the definition of it?

Whether or not the various units are integrated or differentiated depends on a prior decision to conceptualize – sometimes they are unified into a single concept. Whether or not there is differentiation depends on the cognitive benefit of conventionalizing a difference.

But how does it different from nominalists idea of convinience? And how do we determine if there is a “cognitive benefit” in conceptualization?

For instance, “red” refers to a range of wavelengths, meaning “any in that range”, but nevertheless it must be within that range.

But how is “that range” determined? On what objective grounds?

Thomas, thank you for answers! Have I understood you correctly that the difference between attributes and different quantities of one attribute is just perceptually given?

5

Thomas, thamk you for answers. Have I understood you correctly that the difference between different attributes and different quantities of the same attribute is just perceptually given?

6

But then, what is attribute? What’s the definition of it?

Attributes are self-evident. They are sensory data. Even animals can see attributes. The prey of a tiger can see ‘Orange, black stripes’. The difference between us and them though, is that we can conciously identify them, apart from the physical object (i.e. within our heads) whilst the animal can only make this recognition and hold the immediate percept

In fact, it’s practically impossible not to conciously identify attributes when first forming a concept of something perceptual, since the recognition of most kinds of attributes (e.g. Shape, Colour, Distance) is done very early on in life.

I say practically impossible, because there is no way a child can develop a kind of philosophy that teaches man’s mind is impotent and concept formation is useless, before he has developed his basic conceptual faculty.

(If any of that is unclear, I apologise, I’m still learning the finer points of ITOE.)

7

But how does it different from nominalists idea of convinience? And how do we determine if there is a “cognitive benefit” in conceptualization?

I don’t know what their idea is – I don’t find comparative philosophy interesting. As for what is sufficient grounds for conceptualization, I don’t think there can be a general description that doesn’t say “when it is sufficiently useful”. It is no doubt related to how often the X is relevant to people’s lives, how hard it is to distinguish the X’s from the similar Y’s, and how long a compositional phrase is needed to correctly describe X as distinct from Y.

But how is “that range” determined?

It relates to energy / wavelength absorption by the cones in your eye, which I guess reduces to something about quantum mechanics. Essentially, reddish things are physically similar and we react to that similarity in a similar way.

8

Eurgh, re-reading what I said, I see it isn’t really clear at all.

What I meant, essentially, is that ‘attributes’ are initially to do with perceptual reality. They aren’t imaginary things we make up for the sake of convenience. They are, initially, perceptual things with no significance, beyond their own existence. They’re just certain wavelengths of light or sound, or what have you.

They become actual ‘attributes’, in the philosophically meaningful sense (i.e. as conceptual sub-components), when a reasoning mind identifies them as distinct from one another and gives them their own conceptual identity. Though we already automatically identify sensory data, just like animals, it is up to the use of our conceptual faculty to interpret that data into distinct idenities ‘red’, ‘blue’, ‘round’, ‘square’.

9

Thomas, thank you for answers. Have I understood you correctly that the difference between different attributes and different quantities of the same attribute is just perceptually given?

Sorry for taking so long to get back to this thread.

The best way to identify what is given in perception and what isn’t is to make observations. We observe entities perceptually insofar as that entity is not too small nor too far away nor too large for us to see the whole thing. For example, as I write this essay, I can see in front of me a computer, a can of Coca-cola, an ashtray with a cigarette burning in it, a pack of cigarettes, some papers, some pens, etc. The important point to realize with these observations is that it does not require volitional cognition in order to observe those entities, as they are given in perception. Their attributes are also given in perception, their color, size, weight, smell, smoothness or roughness, the taste of the cola and the cigarette, etc. It doesn’t require a theory to be aware of these entities and their attributes – we just observe.

But in order to conceptualize them and to mentally isolate out the different attributes does require cognitive functioning on the volitional level.The coke can, for example, is red, round, silvery, cold, heavy, etc. While these are given in perception, in order to form the concepts “red,” “round,” “silvery,” “cold,” “heavy” etc. requires the full use of our volitional cognitive abilities to mentally isolate out those attributes and to conceptualize them. In other words, the color red is given in perception, but considering it as an attribute and conceptualizing it is not given in perception.

The holdover a lot of people have without realizing it comes from Kant’s assertion that conceptualization is involved in observing something and its attributes. In other words, according to him, there are no givens of observation, but rather it is mental functioning that is involved in making observations. Objectivism draws a sharp distinction between physiological functioning – i.e. perception – and cognitive functioning – i.e. conceptualization. In other words, the human mind qua consciousness does not create the entities nor their attributes, but the human mind does create concepts. Not having that distinction can lead to all sorts of confusions.

While cognition, in the sense of being aware of existence, starts with the perceptually given, it is not your volitional conscious mind that creates those observations. The same can be said for introspection, which is an observation of what one’s mind is doing. What you observe is the given, and then you can mentally isolate out various aspects of that observation.

10

I don’t know what their idea is – I don’t find comparative philosophy interesting. As for what is sufficient grounds for conceptualization, I don’t think there can be a general description that doesn’t say “when it is sufficiently useful”. It is no doubt related to how often the X is relevant to people’s lives, how hard it is to distinguish the X’s from the similar Y’s, and how long a compositional phrase is needed to correctly describe X as distinct from Y.It relates to energy / wavelength absorption by the cones in your eye, which I guess reduces to something about quantum mechanics. Essentially, reddish things are physically similar and we react to that similarity in a similar way.

If you look at a color spectrum, light blue and dark blue really arent any more similar than purple and red. We dont group light and dark blue together because they are similar (they arent) or because they share some common universal of ‘blueness’ - we group them like that because we were taught to do so as children and thats just how we now think. Some other cultures have totally different ways of grouping colors and may not recognise light and dark blue as being ‘shades’ of the same color at all (any more than we would call purple a shade of red). Theres not really any strong objective basis for saying that light and dark blue are the ‘same’ color while green is a different one, since a look at a spectrum may well show that the dark blue is actually more similar to shades of green than it is to the light blue.

Color classifications have a lot of arbitrariness to them and are largely perceptual rather than conceptual/functional - I dont think they are a very good model for concept formation in general. Its a bit of a special case.

The holdover a lot of people have without realizing it comes from Kant’s assertion that conceptualization is involved in observing something and its attributes. In other words, according to him, there are no givens of observation, but rather it is mental functioning that is involved in making observations.

This is ambiguous: in a sense, Kant believed there was an absolute distinction between ‘raw’ unprocessed observations (intuitions, in his terminology) and the concepts that are formed from them. Its probably the most fundamental aspect of his epistemology - most of the first half of the CPR is spent talking about how the ‘givens’ (‘manifold of perception’) is synthesised into intelligible objects.

While cognition, in the sense of being aware of existence, starts with the perceptually given, it is not your volitional conscious mind that creates those observations.

This is Kant’s position and he would have agreed with you.

11

If you look at a color spectrum, light blue and dark blue really arent any more similar than purple and red.

Well, let’s see. Blue is in the range 440–490 nm and red is 625-740 nm. Violet is in the range 380–420 nm whereas purple is a mix of red and blue. If you meant to be comparing the pair (violet vs. blue) and (red vs. blue) then there’s a clear greater physical similarity between blue and violet than between red and violet – and even moreso within the range blue. The physical similarity of the blues explains why they are perceptually similar. OTOH if you did mean the non-spectral color purple then the physical difference is even more pronounced. But I think it’s a mistake to only consider linear wavelength relations, since humans don’t perceive colors by a continuous color-detecting organ, so the comparison ought to be based on the different sensitivities of the S, M and L cones in terms of peak and bandwidth, which determines the nature of what the mind actually sees. The M and L sensitivity graphs overlap a fair amount, whereas S and L are pretty well distinguished, which of course explains why red and blue are totally different.

We dont group light and dark blue together because they are similar (they arent) or because they share some common universal of ‘blueness’ - we group them like that because we were taught to do so as children and thats just how we now think.

That’s partly true and partly not true. Blues are certainly not blues because of a “universal”, since there are no “universals” (“universals” don’t exist), and children are not taught “blue”, they learn “blue” (inductively and on their own). To some extent, people actually do have to be explicitly instructed in color terminology, but that’s secondary. For example, the color “olive” is further hair-splitting from the general concept “green” (acquired by normal lnaguage acquisition) which is quite optional for all cultures. On the other hand, “green” versus “yellow” is not optional for speakers of English, nor are “red” versus “brown”. The right way to understand the relationship between colors and physical fact is to look at the optional splits, to see whether there is a hierarchy of distinctions vs. unifications, and then look at the physical basis of that distinction – and there is a physical basis.

Some other cultures have totally different ways of grouping colors and may not recognise light and dark blue as being ‘shades’ of the same color at all (any more than we would call purple a shade of red).

I don’t understand which claim you’re making. If you’re saying that there is a culture which does not have a unified concept blue and instead actually has unrelated “light-blue” and “dark-blue” concepts, then I’d like to see the evidence for that claim. Name the culture and the experiment, for example. But if you’re saying that there is a cultural aspect to where the boundaries are drawn between what’s in color concept A vs B, then of course, that’s what Objectivist epistemology says. Though Rand didn’t develop a theory of color perception (naturally), the basic elements of Objectivism can and must be applied to the question, in order to understand why color concepts are what they are. Those basic elements are “reality” and “focus”. “Reality” in this case refers to the physical properties of light, the physiological properties of the eye, and all sorts of complex and non-optional visual processing in V1 and so on. “Focus” refers to the fact that differentiation of the perceptually similar requires focus, so “dark” and “bright” can be further differentiated by focusing on the fact that some things do look different from simple “dark” and “light”, giving rise to “red”, and so on. Thus the physically very similar “pink” and “orange” arise from successive focusings on detectable differences within the “red” range, so on Earth, you will not find color systems that minutely focus on that difference while completely ignoring the green-blue and blue-yellow difference.