Ed Buckner raises this question, and he wants my help with it. How can I refuse? I'll say a little now, and perhaps more later.
Kant was brought up a rationalist within the Wolffian school, but then along came David Hume who awoke him from his dogmatic slumber. This awakening begins his Critical period in which he struggles mightily to find a via media between rationalism and empiricism. The result of his struggle, the Critical philosophy, is of great historical significance but is also an unstable tissue of irresolvable tensions. As a result there are competing interpretations of his doctrines.
I will propose two readings relevant to Ed's question. But first a reformulation and exfoliation of the question.
Can one think about God and meaningfully predicate properties of him? For example, can one meaningfully say of God that he exists, is omnipotent, and is the cause of the existence of the natural world? Or is it rather the case that such assertions are meaningless and that the category of causality, for example, has a meaningful application only within the realm of phenomena but not between the phenomenal realm as a whole and a putative transcendent causa prima? Are the bounds of sensibility (Sinnlichkeit) also the bounds of sense (Sinn), or are there senseful, meaningful assertions that transgress the bounds of sensibility?
Weak or Moderate Reading. On this reading, we can think about God and meaningfully make predications of him, but we cannot have any knowledge of God and his attributes. We cannot have knowledge of God because knowledge necessarily involves the interplay of two very different factors, conceptual interpretation via the categories of the understanding, and sensory givenness. God, however, is not given to the senses, outer or inner. In Kantian jargon, there is no intuition, keine Anschauung, of God. All intuition is sensible intuition. The Sage of Koenigsberg will not countenance any mystical intuition, any Platonic or Plotinian visio intellectualis, at least not in this life. That sort of thing he dismisses in the Enlightenment manner as Schwaermerei, 'enthusiasm' in an obsolete 18th century sense of the English term.
But while Kant denies that there is knowledge of God here below whether by pure reason or by mystical intuition, he aims to secure a 'safe space' for faith: "I have found it necessary to deny knowledge in order to make room for faith." (Preface to 2nd ed. of Kritik der reinen Vernunft, 1787, B xxx.) Now if God and the soul are objects of faith, this would imply that we can think of them and thus refer to them even if we cannot have knowledge of them.
The soul is the object of the branch of metaphysica specialis called rational psychology. Since all our intuition is sensible, there is no sensible intuition of the soul. As is well-known, Kant denies that special metaphysics in all three branches (psychology, cosmology, and theology) is possible as science, als Wissenschaft. To be science it would have to include synthetic a priori judgments, but these are possible only with respect to phenomena.
Kant's key question is: How are synthetic a priori judgments possible? He believes they are actual in mathematics and physics, and would have to be actual in metaphysics if the latter were a science. To put it quick and dirty: synthetic a priori judgments are possible in math and physics because the phenomenal world is our construction. The dignity and necessity of the synthetic causal principle -- every event has a cause -- is rescued from the jaws of Humean skepticism, but the price is high: the only world we can know is the world of phenomena. Things in themselves (noumena in the negative sense) are beyond our ken. And yet we must posit them since the appearances are appearances of something (obj. gen.). This restriction of human knowledge to the physical rules out any knowledge of the metaphysical.
On the moderate reading, then, Kant restricts the cognitive employment of the categories of the understanding to phenomena but not their thinking employment. We can think about and refer to the positive noumena, God, the soul, and the world as a whole, but we cannot have any knowledge of them. (And the same goes for the negative noumena that correspond to sensible appearances.) We can talk sense about God and the soul, and predicate properties of these entities, but we cannot come to have knowledge of them. Thus we can meaningfully speak of the soul as a simple substance which remains numerically self-same over time and through its changing states, but we cannot know that it has these properties.
The arguments against the traditional soul substance of the rationalists are in the Paralogisms section of KdrV, and they are extremely interesting.
Strong or Extreme Reading. On this reading, we cannot talk sense about positive or negative noumena: such categories as substance and causality cannot be meaningfully applied beyond the bounds of sensibility. Riffing on P. F. Strawson one could say that on the strong reading the bounds of sensibility are the bounds of sense. This reading wins the day in post-Kantian philosophy. Fichte liquidates the Ding an sich, the neo-Kantians reduce the transcendental ego to a mere concept (Rickert, e.g.), the categories which for Kant were ahistorical and fixed become historicized and relativized, and we end up with a conceptual relativism which fuels a lot of the nonsense of the present day, e.g., race and sex are social constructs, etc.
How's that for bloggity-blog quick and dirty?
So my answer to Ed Buckner's title question is: It depends. It depends on whether we read Kant in the weak way or in the strong way.