Many different kinds of atomism were available in the 17th century. Gassendi and followers like Walter Charleton argued for a roughly mechanist version of atomism on which atoms had only the qualities size, shape, solidity and motion, and all other qualities of bodies supervened on the qualities of atoms.1 Others opted for Galileo’s mathematical point-atoms; atoms endowed with forms; atoms for each of the distinct elements, whether chemical or Aristotelian; and so on.2
One can fairly easily reconstruct something like the primary quality – secondary quality distinction for each of these various forms of atomism. If you treat the determinable properties whose determinants atoms possess as the primary qualities, then everything else is a secondary quality that must derive from those primary qualities. Thus someone who thinks there are four types of atoms, one for each of the four Aristotelian elements, would identify the primary qualities as heat, cold, wetness and dryness, and treat all other qualities as secondary. Indeed, Gassendi uses the term ‘primary quality’ as a synonym for ‘elemental quality’ in precisely this way (1.494a).3
Here, however, I shall limit my discussion to the primary quality-secondary quality distinction as it arises in mechanical atomism. Limitations of space preclude examining how all the different forms of 17th century atomism treat qualities. Moreover, we begin to lose our grip on the primary quality-secondary quality distinction quite quickly once we step away from roughly mechanist philosophies. As philosophers now commonly think of it, the primary quality-secondary quality distinction has a number of components. Both primary and secondary qualities are types of sensible qualities.4 However, they are ontologically – not merely epistemically – different in some important way. The primary qualities are size, shape, solidity and motion. The secondary qualities are color, taste, smell and the like. The primary qualities are used in scientific explanation; the secondary qualities are not.5 Abandoning too many of these components makes us lose our grip on the distinction.6 Mechanical atomism is the only form of atomism for which a primary quality-secondary quality distinction with these components is even a contender, so I shall limit my discussion to it.7
However, discussing the primary quality-secondary quality distinction in mechanical atomism is a somewhat odd thing to do. Gassendi does not explicitly distinguish primary and secondary qualities, and I shall argue that none of the various ways he carves up qualities amounts to such a distinction. One of my goals is to clarify the ontology Gassendi assigns to the qualities that are paradigmatically thought of as secondary. This will show that they are more like primary qualities, as traditionally understood, than secondary qualities. The lack of a primary quality-secondary quality distinction in Gassendi may seem somewhat surprising to 21st century readers. For it strikes us as a very natural distinction for a mechanical atomist to draw. Thus, my second main goal is to explain why Gassendi draws the distinctions between qualities he does, and why he does not draw any clear distinction between the traditional primary and secondary qualities.
I shall proceed as follows. I begin, in (II), with the problem atomists face. Since atoms lack all qualities except size, shape, solidity and motion, we must explain the various macro-level qualities as categorical features deriving from the configuration or arrangement of atoms; as powers; or as appearances in the mind. When Gassendi speaks of the configuration or arrangement of atoms, he typically speaks of the contextura – the way in which atoms are woven together.8 I translate this term as ‘texture’. For although this is an archaic usage, Locke and Boyle used ‘texture’ to speak of the arrangement of corpuscular parts.9
An important question is which of these three views Gassendi adopted and why. Contemporaries such as Descartes and Galileo held that primary qualities are categorical features of bodies while secondary qualities exist only in the mind. A generation later, Boyle and Locke agreed that primary qualities are categorical features but argued that secondary qualities are better thought of as powers. The texture view, the power view and the appearance view were all live options for Gassendi.
In (III), I outline the various distinctions Gassendi draws between types of qualities, and in (IV), I look at his overarching argument that qualities are modes of bodies and consider why he rejects the Galilean appearance view. I consider his reasons for favoring the texture view over the power view in (V). In (VI), I consider a puzzling passage that seems to suggest the Lockean claim that there is nothing in bodies resembling the heat we feel. This is a problem for my interpretation because it is naturally read as contrasting ideas of color and heat with ideas of shape and solidity, which do resemble something in bodies. In (VII), I draw on the distinction between appearance and reality and on Gassendi’s Epicurean claim that the appearances are always true to show that he draws no such contrast. In (VIII), I argue that because he holds that qualities like color and heat are textures, Gassendi in effect rejects the primary quality-secondary quality distinction. For an implication of his view is that while there is an important epistemic difference between qualities like size and shape on one hand and qualities like heat and color on the other, there is no real metaphysical difference. Finally, in (IX) I conclude by suggesting a reason why Gassendi may not have distinguished between primary and secondary qualities in the way many of his 17th century contemporaries did.
17th century philosophers were familiar with the description of Democritean atomism provided by Diogenes Laertius:
The first principles of the universe are atoms and empty space … the atoms are unlimited in size and number … and thereby generate all composite things … The qualities of things exist merely by convention; in nature there is nothing but atoms and void space.10 Not all scholars agree that ‘by convention’ is an accurate translation of Diogenes’ Greek. But in his Physiologia Epicuro-Gassendo-Charletonia, Walter translates the relevant part into Latin as follows: “Lege enim color, lege amaror, lege dulcor” – “For color exists by law, bitterness by law, sweetness by law.”11 He goes on to explain that
[Q]ualities … [ought] to be reputed not as absolute and entire Realities, but simple and occasional Appearances, whose specification consisteth in a certain modification of the First Matter, respective to that distinct Affection they introduce into this or that particular sense, when thereby actually deprehended. Not that Democritus meant, in a litteral sense, that their production was determinable ex institutio hominum, by the opinionative laws of mans Will … but in a Metaphorical [sense], that as the justice, injustice, decency, turpitude, culpability, laudability of Human actions, are determined by the Conformity or Difformity they bear to the Constitutions Civil, or Laws generally admitted, so likewise do the whiteness, blackness, sweetness, bitterness, heat or cold, of all Natural Concretions receive their distinct essence, or determination from certain positions and regular ordinations of Atoms.12
Set aside how odd it is to find an early modern philosopher suggesting that moral qualities are qualities human actions bear in relation to the civil law. Another oddity is more significant: in this brief passage, Charleton appears to suggest three different ontologies of the qualities of macroscopic bodies: