As a teenager he was greatly interested in classics, English literature, poetry, and science fiction.

Williams was awarded a Sheldon Traveling Fellowship and studied philosophy in France and Germany, 1928–29.

He spent the rest of his career at Harvard, arguing for realism, constructing a trope ontology, and advocating a four-dimensionalist theory of time, among many other things such as defending the legitimacy of inductive inference.

As a teacher he lectured for several decades mostly on metaphysics and the history of philosophy, influencing many students who went on to have successful careers and make an impact on analytic philosophy such as Roderick Chisholm, Nicholas Wolterstorff, and David Lewis.

Williams thinks the goal of empiricism is to describe and explain the 'foreground of experience' and to 'intelligibly and credibly construct our account of the rest of the world'.

[4] The traditional understanding of empiricism, going back to John Locke, accepts the foreground of experience as the ground from which we construct our concepts and confirm certain conclusions that inform us of the rest of the world, especially other parts that are not experienced.

He describes empirical realism as follows: Our "empirical realism" is realistic in the most venerable sense, that it affirms there is valid knowledge of the nature and existence of a world distinct from our perceptions and independent of them, and a fortiori distinct from and independent of our thought and speech about it, and yet empiricistic in the classical sense that it affirms that all knowledge, including this, consists of conceptual constructions collected from and confirmed by sensory experience.

Thus (unless there is definite evidence to the contrary) the population probably approximately matches the sample in proportion.

Another benefit is that his theory explains facts about predication without positing substances as members of a primitive category of being.

The main argument for trope theory is that it is the most ontically parsimonious hypothesis with the greatest explanatory power.

Armstrong notes, this distinction between analytic ontology and speculative cosmology is a division within metaphysics.

He argues that any appeal to temporal experience or a direct phenomenological intuition of time's passage is bogus.

[21] Any sense we have of time's passage can be explained in terms of the B-theoretic distribution of content in the four-dimensional manifold.

[22] The crux of his hypothesis is that facts about the B-theoretical distribution of content at the fundamental level of the four-dimensional manifold can do the necessary work in our explanation of the passage and arrow of time.

During this period the type of metaphysics he pursued was unpopular and ridiculed by logical positivism, ordinary language philosophy, and the later Wittgenstein.

In addition, he fought back on various occasions – first against logical positivism and its verificationist theory of meaning and conventionalism about the a priori,[23] and second against Wittgensteinian critiques of metaphysics.

[24] In other places he just scoffed at the ordinary language suggestion that problems about the world can be dissolved by studying the meaning and use of words.

He once wrote: 'It is hard enough to keep one's face straight at any of those readings of philosophical fortunes from the tea leaves of language which our British cousins call "logic" nowadays, but when one knows that the very leaves have been rigged, it is time to lodge a protest, good manners be hanged'.

In the late 1960s and 1970s his work was studied closely at the University of Sydney by such philosophers as D. M. Armstrong, John Bacon, Keith Campbell, and David Stove.

Armstrong writes: 'It so happens that Sydney University was for some years the world centre of Donald Williams studies'.

[26] Armstrong appreciated and implemented Williams's conception of metaphysics and his distinction between analytic ontology and speculative cosmology in his theory of universals.

[27] Bacon and Campbell embraced Williams's trope theory, expanding on the ontology and ensuring it remained a strong contender in the literature.

[30] His instruction at Harvard spans several decades and he came into contact with many students who went on to have respectable careers in philosophy, e.g., Roderick Chisholm and Donald Davidson.

He supervised Nicholas Wolterstorff whose theory of universals as kinds bears a strong resemblance to Williams's ontology.

Lewis owes in part his Humean supervenience, four-dimensionalism, and metaphysics of time more generally to Williams.

For Lewis, the way out of the sceptical worry was to take actuality as a relative matter and interpret the word 'actual' as an indexical.

[33] Later on his career Lewis gravitated more and more towards Williams's ontology, expressing sympathy with a sparse theory of tropes.

Armstrong and Lewis were both influential figures who played their own part in bringing metaphysics back into mainstream philosophy.

Williams's articles on trope theory and four-dimensionalism are classic articles in analytic philosophy – anthologised in many places and recently collected together with previously unpublished essays that advance his metaphysical theory in The Elements and Patterns of Being: Essays in Metaphysics (2018).