This is a piece on old-style rationalism. However, it will still – at least to a small degree – apply to any contemporary rationalist positions. (For example, those of Laurence BonJour, James Van Cleeve, etc.)
Descartes is a good place to start:
“Descartes thought that only knowledge of eternal truths – including the truths of mathematics, and the epistemological and metaphysical foundations of the sciences – could be attained by reason alone; other knowledge, the knowledge of physics, required experience of the world, aided by the scientific method... Descartes developed a method to attain truths according to which nothing that cannot be recognised by the intellect (or reason) can be classified as knowledge. These truths are gained 'without any sensory experience', according to Descartes.”
The very idea of "thought" existing independently of all experience seems bizarre. It simply doesn’t sound feasible at an intuitive or prima facie level. For a start, the words used by the rationalist will have been learned from experience. The logical rules and inferences, likewise. Even if what the logical rules and words express and not themselves part of experience, this doesn’t stop it from being the case that the words for these concepts, propositions and rules were learnt from experience. One could say that it's no longer preaching from the top of a rationalist tower if empirical and experiential means helped the rationalist get to the top of that tower.
Of course not all rationalists have denied these pollutions of their pure rationalism. Indeed it would be silly to do so.
And what, exactly, would the pure rationalist think about? Does he always think about some Platonic realm of abstract objects? Does he think about propositions and other abstract entities? When he thinks about truth and meaning, does he think entirely of these in the “pure light of reason”: a light that's untouched by empirical vicissitudes and contingencies?
Of course thought is more than experience. That is, experience is experience and thought is thought. However, thought relies on experience. Or, in Kant’s philosophy, thought relies on “sense impressions” (or “phenomena”) in order to get going. The actual things that the mind does with these experiences or sense impressions aren't themselves experiences or sense impressions. A logical inference, for example, isn't a sense impression or an experience. In addition, making a generalisation about experiences or observations depends on the experiences or observations. However, the generalisation itself isn't an experience or an observation. Indeed mental images or acts of the imagination aren't themselves sense impressions or experiences either. Nonetheless, they too may depend on sense impressions or experiences.
So not many empiricists would claim that thought and experience are identical. Yet the hard-core rationalist appears to dispense entirely with sense experience or impressions - even if only philosophically. The empiricist doesn't - and can't - possibly dispense with thought. Of course the very articulation of an empiricist position requires thought. Not only thought: thought that's not entirely dependent on sense impressions or experiences.
Rationalists have said that “thought is the only source of knowledge”. Admittedly, some empiricists have also said that experience and sense impressions alone account for all knowledge. Others have said that all knowledge is “dependent” on experience or sense impressions.
What gives the rationalist so much confidence or faith in their rationalism? Traditionally, it's usually been the status and existence of necessary truths. That is, rationalists have argued that empiricists can't give us necessary truths. In fact empiricists admit that they can't give us necessary truths.
So why are these necessary truths so special to the rationalist?
They're special because they can be known to be both necessary and true without recourse to experience or sense impressions. And that's precisely what the rationalist requires of truth – certainty and necessity. However, even though these necessary truths aren't reliant on experience or sense impressions, they're still at times necessary truths about the world. That is, they're things that must be the case about the world. Indeed the world can't be any other way than that expressed by such necessary truths. Having said that, when we observe the world we don't discover these necessary truths about the world. They (as it were) come before observation. In certain cases (e.g., Kant's), they're a “necessary condition” for observing or experiencing the world. So these necessary truths are applicable to the nature of the world; though aren't derivable from the world. In that sense, these necessary truths about the world must transcend the world. They transcend the world as it's observed or experienced. More clearly, certainties and necessities are never experienced at all. They're known exclusively through thought. Experience can't tell us (or help us) when it comes to discovering (if discovering is an appropriate word here) these certainties and necessities.
And all this is precisely why empiricists have rejected the very notions of necessity and certainty. Such things can't be experienced or observed. We don’t get sense impressions from things that are necessary or certain. That's because necessity and certainty aren't empirical things to get sense impressions from – not even when they belong to things or conditions which aren't themselves necessary or certain.
In addition, according to certain empiricists, a world with necessities and certainties would be indistinguishable from a world without necessities and certainties. In other words, it's is a difference which doesn’t make a difference.