Transcendental idealism, whether it is an epistemological thesis or not, has obviously many epistemological implications, such as the need for epistemic humility resulting from the ignorance of things in themselves. However, whether the distinction between noumena and phenomena is merely an epistemological or methodological distinction or a more substantive metaphysical doctrine is a much-debated issue. The debate regarding the noumena/phenomena relation is often phrased in terms of the number of worlds that result from particular theories, so that epistemological interpretations are seen as `one-world' views, whereas metaphysical interpretations are considered to be `two-world' views. This classification is rather inadequate and it is better to differentiate theories on the basis of whether or not they are epistemological or metaphysical in nature, since this distinction is more fundamental. A fully adequate classification, however, must classify interpretations along both dimensions since they run orthogonal to each other.
The distinction between noumena and phenomena can either be seen as a metaphysical or as an epistemological distinction. According to metaphysical interpretations, the distinction between noumena and phenomena is a distinction that derives from the objects. It is not based on the way we consider or conceive of the objects. Instead, there is some metaphysical distinction or difference that pertains to the objects and which lies at the basis of the noumena/phenomena distinction. Proponents of the epistemological or methodological interpretation claim that the distinction is a distinction between ways of considering the same objects. They claim that we view it from different viewpoints, that in thinking about them we occupy different standpoints, namely the transcendental and empirical standpoints. When we consider things as they are in themselves we are abstracting from the conditions of our sensibility. Thus, the terms `noumena' and `phenomena' refer to different entities or metaphysical features of entities, if one takes the former approach, while they refer to different ways of conceiving, considering or knowing entities, according to the latter approach.
Interpretations of transcendental idealism are usually distinguished by assessing the number of worlds that they imply. This criterion is slightly problematic in that it might be better to talk about domains of discourse, rather than about `worlds'. This is because there are positions, such as the `virtual object' view espoused by van Cleve, that do not really classify as two-worlds views even though there are two separate domains. In the case of van Cleve's interpretation there is only one world, but there are two ontological domains or domains of discourse. There are two sets of entities, but only one set consists of real existents, whereas the other's members are virtual objects. As long as we have a very weak understanding of worlds, i.e. no substantial ontological reading of worlds, these will amount to the same. (Though the terminology of `worlds' is slightly misleading, it has become so entrenched in the literature that I will stick to it, with the proviso that one should keep in mind that a weak reading of worlds is used that does not have any serious ontological import.)
The metaphysical/epistemological and the one-world/two-worlds distinctions constitute two orthogonal distinctions that can be combined, giving us four possible groups of interpretations of the distinction between noumena and phenomena. Metaphysical interpretations can be classified as one-world or two-world theories and the same applies to epistemological approaches. These four options are illustrated in the diagram below:
Metaphysical interpretations can be either one-world theories or two-world theories. Usually, they imply the existence of two-worlds, but there are also some well-known metaphysical one-world approaches. The most straightforward metaphysical view is that noumena and phenomena are different objects. They are ontologically distinct entities. This option then, of course, brings with it the well-known two-worlds view. Within this broad category of metaphysical two-worlds views, there is a vast range of more particular views that differ in the account they give of the nature of noumenal objects and phenomenal objects, as well as the relation that exists between them.
If the distinction between noumena and phenomena is a metaphysical distinction, then this implies that noumena and phenomena must be distinct, but it does not mean that we have distinct objects and distinct worlds occupied by these different objects. Noumena and phenomena can be distinct aspects or parts of the same object. For example, Warren and Langton argue that there is one object, but that noumena are collections of intrinsic properties of objects, whereas phenomena are collections of the relational properties of the same objects. On this view, phenomena and noumena are not distinct objects, but rather refer to bundles or sets of properties. We can talk about the `phenomenal object', as long as this `object' is not understood as a mereological unity, as an object in the full-blown sense. Only noumenal objects are objects in the full-blown sense and this explains their ontological priority. So, we have a clear metaphysical distinction between two kinds of properties.
There are several other candidate distinctions that can be appealed to in order to arrive at metaphysical one-world views. Possible options include views according to which Kant was concerned with different essences, or that the distinction between noumena and phenomena is simply the distinction between substrata and their properties. A further one-world possibility is to treat it as a distinction between different mereological parts of objects (whereby the notion of parthood has to be understood in a non-spatial way, i.e. no reference to overlap etc.).
Epistemological interpretations can also be combined with one-world and two-world ontologies. That is, the distinction between noumena and phenomena is a distinction between ways of considering or conceiving of entities and this understanding can be combined with a belief in the existence of different ontological domains. After all, it is an epistemological or methodological issue and as such should not decide the metaphysical question how many worlds there are. Usually, epistemological interpretations are part of one-world theories since the main motivation for accepting an epistemological understanding of the noumena/phenomena distinction is to avoid the `crazy' metaphysics of transcendental idealism. While they are mostly one world views, it is clearly possible to combine them with two-worlds, for example in order to account for Kant's metaphysical claims that are based on ethical considerations regarding freedom, God and immortality. In this case the two worlds will not be described as the noumenal world and the phenomenal world since the terms `noumenal' and `phenomenal' describe the ways a world can be considered. Instead, the worlds can be described as the transcendental world and the transcendent world, whereby the later one includes God and similar objects that are beyond the bounds of experience. The former world is a world of things that can be regarded as phenomena or as noumena, as they appear to us or as they are in themselves, whereas the the later world is a world in which there are things that do not appear to us and can only be considered as they are in themselves. Thus, the epistemological interpretation implies that there is one world of entities that can be considered as phenomena and as noumena, but this does not imply that there cannot be other worlds inhabited by objects that can only be considered as noumena. While this latter option does not have much appeal since epistemological readings are used to dispose of `outrageous' metaphysical claims, it seems nonetheless to be a coherent option.
Thus it seems that there are two different dimensions along which interpretations of the noumena/phenomena distinction can vary. The number of worlds criterion runs orthogonal to the criterion regarding the nature of the noumena/phenomena distinction. There are disputes along both dimensions. However, the dispute regarding the nature of the distinction appears to be more fundamental. There is much more similarity within metaphysical approaches, on the one hand, and epistemological approaches, on the other, than within one-world interpretations or two-world interpretations.