Writing about Leibniz on expression got me thinking about other early modern talk about expression, and in particular about Spinoza, who talks several times in his Ethics about things expressing others. Some of this expressing involves language, but other cases seem not to. Thus both attributes and modes are said to express things. For example, 1p6 talks of the infinite attributes of God, “each of which expresses eternal and infinite essence”. Modes, meanwhile, are also said to express God’s essence (though in a certain way, related to a certain attribute). Thus, Spinoza says in 2d1 that “By ‘body’ I understand a mode that expresses in a definite and determinate way God’s essence in so far as he is considered as an extended thing”, and in 2p1d that “Individual thoughts, or this and that thought, are modes expressing the nature of God in a definite and determinate way”. There is also related language in which attributes are said to be expressed in modes (see 3p6d, and perhaps 1p25c).
My initial question is relatively simple: what does Spinoza mean when he talks of attributes and modes expressing in these ways? I consider this by presenting at two different possible models of expression, and asking whether the expression of attributes and modes fits either model.
Two models of expression
- A causal model of expression. One Leibnizian idea about expression is that “every effect expresses its cause” (Discourse on Metaphysics 28). This helps, I think, understand some otherwise puzzling things Leibniz says about substances expressing God. Might it also help us in understanding what Spinoza says about attributes and modes expressing the essence of God?
- A linguistic model of expression. At various points in the Ethics, definitions are said to express (1p8s2, 1p16p), words are said to express (Explication of 2d3), and people are said to express using words (2p47s; see also 2p40). In these cases, expression seems to be a descriptive or representational relation. These are not all the same case, but they are plausibly related, and use a sense of ‘expression’ we might at least recognize. Could something like this sort of expression be a good model for the metaphysical cases?
Does the causal model help?
Suppose we start with the causal model, and with the expression done by modes. If Spinoza held the causal view of expression, the reason why he would say, e.g., that some particular body expressed “God’s essence in so far as he is considered as an extended thing” would be that God’s essence was the cause of that particular body. Now that might appear to be wrong, because the causal interactions of bodies are with other bodies. However, Spinoza does say that God causes modes: e.g., 1p18 says God is the immanent cause of all things. So we might think of Spinoza as saying that modes express their immanent cause, which is God, because God is their cause. So far, this is consistent with the causal model.
What about attributes though? Can we think of Spinoza’s attributes as expressing God’s essence, which is their cause? (One might imagine a causal hierarchy from essence to attribute to mode, with the things lower down on the hierarchy expressing those higher up.)
Someone might deny that Spinoza could have though this, by arguing that there is not enough of a distinction between the essence and the attributes for one to be the cause of the other – e.g., by identifying the essence with the collection of the attributes. Suppose for now though that there is enough of a distinction between essence and attributes, such that they could stand in a causal relation. Is there any evidence at all that Spinoza thought they did?
I suppose one might try and run an argument from God being the immanent cause of all things. If attributes are things, then they are caused by God. But it seems too easy to deny that attributes are things in the relevant sense. 1p18 comes soon after the claim that “nothing exists except substance and modes” (1p15d). In general, it is hard to see how to sustain the view that the essence of substance is the cause of its attributes. At least, that seems an obvious sticking point for using the causal model.
Next time, something on using the linguistic model.
Stewart Duncan 