Aspects of Meaning
Many theories of meaning have been propounded, each seeming to have some merit. But only one theory can be true, so some have to be rejected—or so we suppose. I will contest this. Things are more complicated, more nuanced. Among the theories defended we have: truth conditions, verification conditions, linguistic use, speaker meaning, hearer uptake (e.g., power to cause beliefs). Are these theories really competitors or might they all be integrated together? I am going to distinguish between what I shall call primary meaning and secondary meaning: as a first approximation, primary meaning is strict and literal linguistic meaning in the public language, while secondary meaning is anything that belongs to the overall significance of a speech act—anything conveyed by an utterance (apart from primary meaning). If you consult a thesaurus under “means”, you will find such synonyms as the following: signifies, connotes, denotes, designates, indicates, expresses, conveys, suggests, alludes to, adumbrates, hints at, intimates, implies, points to, stands for, symbolizes, touches on, mentions, calls to mind, and gestures towards. This is a very capacious list and does not correspond to what we ordinarily think of as strict and literal meaning. Yet it would be true to say that these verbs signify properties or actions that depend on literal meaning—they are not independent of it. A meaning-independent property of a word or sentence would be its phonetic or graphic features—how it sounds when spoken or looks when written down. These have nothing to do with what the word or sentence means (put aside onomatopoeia). Likewise, where a sentence is uttered, or at what time, is meaning-independent. So, I shall say that some aspects or instances of meaning are primary and some are secondary, according as they are comprised in strict and literal meaning or not. To be concrete, let us agree that truth conditions are constitutive of primary meaning (this is just to fix ideas): then all the other theories of meaning I listed can be said to belong to secondary meaning. For example, verification conditions are primary-meaning-dependent secondary meanings. Intuitively, they are consequences or concomitants of primary meaning, derivative, parasitic. A sentence can allude to (“show”) something it doesn’t strictly mean or say—convey it, put it across, communicate it. It is something the speaker and hearer know in knowing the language, but it isn’t part of what the words strictly mean. The structure of the position can be compared to Frege’s distinction between sense, on the one hand, and force and tone, on the other. Force and tone are not strictly part of sense, though they depend on sense for their signifying properties; they are a type of secondary meaning—just not the core of meaning, its fundamental nature or nerve. Sense does not depend on them; it forms their foundation. Similarly, we grasp verification conditions (criteria of assertion) when we understand a sentence, but this grasp is not (we are stipulating) part of our grasp of the sentence’s strict primary meaning—though it is a consequence of that. It is a secondary meaning placed on top of, or set besides, the truth-conditional primary meaning. And the same is true of the other items I listed: use, speaker meaning, and hearer uptake (as in so-called causal theories of meaning—the meaning of a sentence is given by the beliefs its utterance causes). We know these facts about a sentence whose meaning we grasp (unlike certain other properties of the sentence), and they depend on the truth-conditional meaning we grasp, but they are not part of its primary meaning—its core, essential, foundational meaning. They are to primary meaning what (for Frege) tone and force are to sense—secondary, derivative characteristics (not the heart of the matter). One way in which they differ is that they are, in a certain sense, agent-relative: they depend not only on the primary meaning but also on the characteristics of the speaker and hearer. Thus, verification conditions depend on the epistemic capacities of the agent in question; and what one agent can discover may not be discoverable by another agent (consider a blind man). Truth conditions are not agent-relative, but verification conditions are. The same is true of speaker meaning (trivially), language use, and hearer uptake (also trivially). These are all pragmatic in the sense that they concern relations between words and people not just words in themselves (as with syntax and semantics, as classically defined). So, secondary meaning is agent-relative and primary meaning is not. I can best explain this by making an analogy with primary and secondary qualities (I chose my terminology by analogy with the older use of these terms). The primary qualities of an object are not perceiver-relative, but secondary qualities depend on both primary qualities and properties of the perceiver. They have a foot in both camps: they depend on primary qualities but they also depend on the perceptual reactions of perceivers. Hence, they are secondary—a bit second-class, parasitic. Color is not as ontologically robust as shape; its claim to centrality is not as strong. A philosophical extremist might insist that objects don’t have color, that color isn’t really a qualityof objects: it is extraneous, imposed, allowed in by courtesy not by right. Similarly, someone might claim that what I am calling secondary meaning isn‘t really meaning; it’s just associated with meaning, a kind of moocher and hanger-on (parasites are not usually of the same species as what they parasitize). I don’t think the word matters much; what matters is the intimate relationship between strict and literal meaning and other aspects of sentences that convey information, unlike things like sound and location. The question is what the semantic natural kinds are, the fundamental taxonomy; and the secondary cases, both for perceptual objects and language, are sufficiently close to the primary cases that we can justify using the words in the way I am suggesting. That is, colors are qualities of objects and verification conditions (etc.) are a type of meaning—though secondary types. We thus make room for a more inclusive picture of meaning, as we do for the case of perceived objects. This allows us to favor one theory of meaning as central and basic without having to declare the others outright false. For the others have their merits and attractions, their staunch and sincere defenders—they aren’t just completely wrongheaded. We just need to recognize that what we call “meaning” has a wider extension than philosophers have tended to assume. Meaning isn’t the kind of monolithic uniform structure that tradition suggests. This is a kind of Wittgensteinian point: let’s not suppose that only the indicative sentence as used in science deserves to be called meaningful. Let’s not be so anal about language and meaning. Meaning is a mixed bag, a ragbag even, not a crystalline Platonic form—a kind of sublime and singular geometry of thought. Objects are a mix of primary and secondary qualities, messy as that seems, and sentences are a mix of primary and secondary meanings, untidy as that may appear. We don’t need to fight with each other over which theory of meaning is correct; we can accept that each theory deserves its place in the sun. Perhaps one theory deserves pride of place over the others, but the other theories are not hopeless losers and misfits with nothing to recommend them. We can be principled semantic pluralists.
 The tendency of most theories of meaning has been to reduce the semantic content of sentences: meaning is only sense or onlyreference or only use. I am suggesting that we increase it dramatically (but not irresponsibly), so that the standard theories get a shot at being admitted to the party. We don’t need to be so selective or snobbish or sniping, though a single type of theory gets to determine the nature of primary core meaning (it isn’t a congeries). The picture is that of a nucleus fanning out to satellites feeding off the energy of the nucleus. Much the same picture fits the case of objects and their qualities, primary and secondary (also tertiary). There is the solid hardworking hub along with its shiftier cloudlike companions. The atmosphere of the earth can be counted as part of the earth, but it isn’t as strictly part of the earth as its rocks and mountains.