In his article “Knowledge by acquaintance and knowledge by description”, Bertrand Russell explores how we can know something about an object even if we lack direct acquaintance with that object. For example, it seems that I can know that the biggest crater on the moon exists without ever being acquainted with that particular crater. In this case, I know something about the crater (that it exists), without being acquainted with that crater in any way. Thus grounds the fundamental distinction between knowledge by acquaintance and knowledge by description. In the case of the largest moon crater, we can have definite and certain knowledge of it (that it exists), without knowing in the slightest anything like where it is located, or how deep it is. Knowledge by acquaintance is a matter of the subject having a “direct cognitive relation” with that object. On Russell’s view, we are not acquainted with physical objects themselves, nor are we acquainted with other people’s minds. Instead, Russell claims that, when it comes to particular things, we are for the most part directly acquainted with sense-data. Although Russell thinks that the most common object to be acquainted with is a sense-datum, he also thinks that we can be acquainted with universals, such as “yellow”, which he calls a “concept”. Russell thus thinks that we are acquainted with two basic types of objects: particulars and universals. The basic particulars which we are acquainted with are sense-data, and the universals we are acquainted with are various kinds of abstract concepts such as “roundness”. In addition to knowledge by acquaintance, Russell thinks there is also knowledge by description. For Russell, a description is generally any phrase such as “the so-and-so” e.g. “the biggest crater on the moon”. And accordingly, we can have knowledge by description of an object without being directly acquainted with that object insofar as we can have knowledge that there exists the biggest crater on the moon without being directly acquainted with it, or know anything about its particular details.
An important corollary of Russell’s theory is that any description of a particular must, ultimately, be cashed out in terms of acquaintance with sense-data. The fact that I know “The biggest moon crater is 13km deep” has to be cashed out in terms of a sense-datum that I have been acquainted with, be that a sense-datum of reading about the moon crater on a website, or the sense-datum of talking to my astronaut friend who visited the moon. And likewise, the knowledge by acquaintance of the biggest crater by the astronaut has to be cashed out in terms of his acquaintance with the sense-data of his exploration of the crater, or the sense-data of his looking through a telescope. In the case of particulars then, Russell is committed to a strong reductionism whereby any knowledge by description of a particular can, in principle, be reduced to knowledge of sense-data. We thus come to what Russell calls a fundamental epistemological principle: “Every proposition which we can understand must be composed wholly of constituents with which we are acquainted” (p. 117). For Russell then, all knowledge of the moon crater as a particular thing (e.g. the fact that it is 13km deep), no matter how abstract, must rest on original first-person acquaintance with the sense-data of the moon. Of course, prior to us being acquainted with the largest moon crater such that we know it is 13km deep, we could have had only some knowledge by description of it, namely, “The biggest crater on the moon exists somewhere, but we’re not sure where yet”.
The epistemological implications of this principle run deep, for it seems like Russell would want to apply it to cases where people claim to have descriptive knowledge of entities that, in principle, no one could have ever been directly acquainted with e.g. God. If we asked what it means to understand the proposition “God exists”, a theist persuaded by the ontological argument might say it requires an understanding of the phrase “an absolutely perfect being exists”, for that is the conclusion of the argument it tries to establish (on the premise that a perfect being is most perfect, and existence is a perfection). But if we were to accept Russell’s epistemological principle, we would need to be acquainted with the concepts of absoluteness and perfection in order to properly understand the proposition “God exists”, for it seems reasonably agreed that theists are not basing their description of God on sense-data of God, except for perhaps schizophrenics. Does Russell’s theory of knowledge prevent us from being acquainted with universals like “perfection”? It’s hard for me to reconstruct what Russell’s actual view is on this question, for we would need to know more about Russell’s views on how we learn about universals. He does say, however, that in regard to learning the universal concept of “yellow”, that “Not only are we aware of particular yellows, but if we have seen a sufficient number of yellows and have sufficient intelligence, we are aware of the universal yellow” (p. 111).
If the case of yellow is analogous with the case of perfection, then it seems we would have to have been acquainted with a sufficient number of perfect beings in order to understand the concept of “perfection”. However, it is highly debatable as to whether anyone has ever been acquainted with a perfect being in his or her lifetime. Thus, if the ontological argument seeks to establish the existence of a perfect being, and such an argument requires the prior acquaintance with the concept of perfection in order to understand the conclusion, and it is only possible to learn about perfection through acquaintance with perfect beings, then the ontological argument cannot actually establish what it seeks to establish (namely, that a perfect being exists). In other words, because we must be acquainted with the concept of perfection in order to know what a “perfect being” is (as per Russell’s principle), and Russell’s principle seemingly indicates that we can only learn about such a concept through acquaintance with perfect beings, then the ontological argument cannot possibly go through because it requires that we know in advance what it sets out to prove, namely, that there exists a perfect being through which we learned about the concept of perfection. But since it seems plausible to suppose that we have no direct acquaintance with perfect beings as finite creatures, then we cannot learn about the concept of perfection in the way theists require, and thus we do not really understand (and thus know) the phrase “a perfect being exists”. That is, if we can not understand that phrase without being acquainted with a perfect being, and the only reason we would have for thinking a perfect being exists is the ontological argument, then it is plain that the argument does not work, for it assumes that one has a prior acquaintance with perfection. But as we have seen, the possibility of having an acquaintance with perfection is what it sets out to prove! One would have to show that you can learn about the concept of perfection without ever being acquainted with perfect beings. This might be possible, but if it were so, then it would be of no help to the theist using the ontological argument, for any supposed jump from having a conception of a perfect being to there actually being a perfect being wouldn’t necessarily follow if we establish that we can learn about perfection without there actually being any perfect beings.