New Apps has a thread on 'great lines in philosophy.' Here's a page of such lines compiled at Siris in 2009. The New Apps thread was occasioned by Jim Holt's NY Times piece on literary style in philosophy. Brand Blanshard gave a talk on philosophical style in 1953. He mentions Mill and Locke as two good stylists.
Blanshard himself wrote well at least sometimes (I haven't read that much of his work). The following passage by Blanshard is too lengthy to count as a great line, but I like the way he here contrasts the rational (or normative) and causal (or mechanical) orders. I also like his use of 'supervenes' in 1966 (though I'm not sure of that date) to capture something about the mind-body relation before that term caught on in philosophy of mind. Donald Davidson is often credited with introducing the term into the philosophy of mind in 1970. While Davidson spoke of mental events as supervening on physical ones, Blanshard writes here of one type of 'law' (or perhaps the 'operation' of one law) as supervening on another.
Thinking in art and morals and even mathematics is neither the reflection in consciousness of a mechanical order in the brain nor the tracing with the mind’s eye of some empirical order in its object, but an endeavour to realize in thought an ideal order which would satisfy an inner demand. The nearer thought comes to its goal, the more it finds itself under constraint by that goal, and dominated in its creative effort by aesthetic or moral or logical relevance. These relations of relevance are not physical or psychological relations. They are normative relations that can enter into the mental current because that current is . . . teleological. Their operation marks the presence of a different type of law, which supervenes upon physical and psychological laws when purpose takes control. (Brand Blanshard, 1966)Actually, I don't know where Blanshard wrote this. I copied this passage from one of his works but noted only the year and not the title of the work in question, and now I can't locate on-line any publication by him in 1966. Hmm. (He also used 'supervene' to characterize relations between different causal orders in his 'Case for Determinism' .)