It is difficult to evaluate the quality of mathematical work or the particular destiny of mathematics in the 20th century. We have already given some criteria: logical, conceptual and didactic clarity and rigor, clarity of intuition (which, contrary to common myths, is not incompatible with logical rigor), an orientation towards synthesis and simplification, an orientation towards applications to the sciences embodying novel unifying ideas, awareness of the deep problems regarding mathematical models of reality, awareness and interest in fundamental philosophical problems.
Logicism is irrefutable. Mathematics is in its purest essence is the deployment of computability (and hence logic), but this game cognitively involves intuition (just as chess strategies) as its develops towards perfection. It is a 'game' in the noblest sense: not 'to calculate' in the sense of mere application, but 'to calculate' as in 'be able to calculate' through finding a strategy and sequence of right moves in a formal game. The very conventionalism is not conventionalist in its a priori logical and computational cognitive presuppositions (see our paper On Analyticity and the A Priori). Moreover, beyond the genius of Frege, Turing and Church (and Brouwer's intuitionism is just computationalism), there is a presupposition in mathematics of a certain objective-intuitive correspondence and specially a claim to enter into the objective truth of the world in the form of science. Kant saw the tip of this iceberg: "Thoughts without intuitions are empty, intuitions without concepts are blind" . The problem of the intuitive correlation of mathematics is similar to that of the efficacy of mathematical models of nature.
We can ask if ancient philosophy was not simply ancient science and mathematics (and its politics and ethics but applied mathematics) or if there is any real difference between the ideals of philosophy, mathematics and science ? Experiment and observations influence the setting up or modification of axiomatic-deductive systems but their computability and logic is ideally the same as that of mathematics and allegedly rigorous philosophy (the appeal to intuition is part of what renders certain arguments plausible).
Mathematics, the science of sciences, occupies a special central elevated region in the geography of the sciences (and the arts and humanities). It could be argued that only mathematics surveys and illumines and contributes to and interprets all the other sciences. And that the other sciences reach their perfection only insofar as they adopt and conform to the language and concepts of mathematics. But mathematics' claim to the throne can be challenged by (philosophical) logic and computation theory, (philosophical) linguistics, (philosophical) psychology and even some other disciplines (recall how we completely lack a formal logic or mathematics of consciousness).
The universalist point of view, the state of mind, corresponding to differential equations and dynamical systems and the direct intuitive awareness of their omnipresence and all permeating nature, is an elevated state of consciousness indeed.
In the 20th century (if the number of mathematical publications increased exponentially) we saw a drastic decrease in logical-conceptual synthesis and clarity and a great lack of philosophical intelligence and awareness, specially regarding the relationship between mathematics and science, regarding the essence and scope of mathematical models themselves (cf. my paper Differential Models, Computability and Beyond).
We saw a proliferation of a kind of junk philosophy and junk science allegedly based on mathematics, specially in the consciousness of the general public, something which can be traced to the reigning ideology of the times (reflected in popular media personalities) and to questionable core elements of 'general systems theory' and 'game theory'. We can name a few of these fads: 'chaos', 'fractals', philosophies of vagueness, randomness and uncertainty, 'bifurcations', certain usages of the terms 'emergent', 'self-organization', 'self-referential' (later we shall discuss the agenda implicit in Smullyan and Martin Gardner as well as Hofstadter's appropriation and distortion of the legacy of Gödel, Escher and Bach - the same concepts the book focused on are deployed with far greater sophistication and artistry in Gaarder's Sophie's World) , 'neural', 'non-linear', 'evolutionary' or 'quantum' beyond its narrow legitimate technical sense and at the same time offering no clear and cogent logical and philosophical account of them.
There are not that many people who could be considered the genuine heirs of Poincaré (in differential geometry, topology and differential equations). These stand out as philosophical-scientific-mathematical giants, towering above others. We have already discussed the problems with the history of algebraic topology and its loss of its geometric and combinatorial essence elsewhere. We name first of all philosopher-mathematician René Thom (and his collaborators and followers) and Stephen Smale (and the Brazilian school of dynamical systems). For Celestial Mechanics Siegel stands out. Jack Hale has written some excellent textbooks. For geometers we have people like Élie Cartan, Georges de Rham, Teichmüller, Leray (father of sheaf theory), Whitney, Morse, Shing-tung Yau, Milnor, Thurston, Roger Penrose, Pierre Deligne, Stanisław Łojasiewicz, Coxeter and John Horton Conway. Ergodic theory (which embodies Lebesgue's measure theory in the mathematical modelling of nature) , originated by Birkhoff, is also important (as is Boltzmann and Shannon). And we attach great importance of fluid dynamics (cf. David Ruelle's theory of turbulence - a distant heir to da Vinci). And also Lyapunov, Pontryagin, A.N. Kolmogorov, A. Fomenko. It may be well that the pioneering contribution of Polish mathematicians like Kuratowski and Sierpiński to topology has been downplayed or ignored (for instance Zorn's lemma was actually proven by Kuratowski and the Hausdorff property was already formulated by him in the 1920s; so-called 'fractals' were already studied by Sierpiński. The Cantor set was discovered previously by Henry Smith in 1874. Indeed the Weierstrass function is a perfect example of a fractal curve. Gauss certainly knew the p-adic numbers and called them congruentia infinita). This list is obviously tentative and incomplete. We also have to show that if Thom does engage in slightly vague discourse at times, at a fundamental level Catastrophe Theory is both mathematical rigorous and philosophically cogent and more generally represents the correct approach to using the methods of differential topology and singularity theory in differential equations and differential mathematical models of nature and linguistics. For profound work in the mathematical philosophy of continuous and smooth structures and their application to physics, Lawvere stands out.