Sören Stenlund, Language and Philosophical Problems, Chapter 2: "Notions of Language and Theories of Meaning"

There are two approaches to the study of meaning: an a priori approach, concerned with the "conceptual" conditions for the possibility of meaning (exemplified by Frege and the early Wittgenstein), and an empirical, naturalistic, approach, which dominates contemporary discussions of meaning. The naturalistic approach, while suitable for certain kinds of limited clarifications, when applied to fundamental questions about the nature of language and meaning, produces confusions. The basic mistake of the empirical approach is to assume that expressions of ordinary language have sense independently of the practices in which they are used. And the practices in which expressions are used are fundamental; they can't be explained in more basic terms. Any attempted explanation will presuppose what it tries to explain.

Pär Segerdahl, Language Use, Chapter 16, "Presuppositions and Methods of Linguistics"

Explaining failures of use by way of presupposition failure is to mistake a linguistic form of description for actual use. Talk of presuppositions makes sense only against a background of normal use. We should aim for a careful description of linguistic practice, not an explanation of it.

Pär Segerdahl, Language Use, Chapter 14, "The Semantic Reading and Indirect Speech"

The notion of the "literal meaning" of a sentence is dependent on applying a particular semantic theory of language to language use. It is a mistake to think that the system of representation that we have for language (a semantic theory, e.g.) is actually in the language itself.

Pär Segerdahl, Language Use, Chapter 13, "Intentions and Beliefs as Conditions for Use"

The picture of language according to which external, lifeless signs need to be supplemented by inner processes, like belief or intention, is a misleading picture. Words have significance in use, and the need for a theory of speaker intentions is occasioned by nothing more than the misleading picture of external, lifeless marks and sounds.

Pär Segerdahl, Language Use, Chapter 12, "Language vs. Languages and Philosophy vs. Linguistics"

The activity of language use cannot be reconstructed using the formal techniques of linguistics and philosophy of language; words only have significance in that they are used in particular activities.

Pär Segerdahl, Language Use, Ch.10, "The Speaker-Hearer Scheme of Communication"

Both the formal semantic approach and the communication-intention (speech-act) approach to the study of language find a distinction between conventional meaning and speaker meaning; that distinction is rejected by Segerdahl. And both approaches involve abstracting away from details fo concrete use. Segerdahl says that a proper appreciation of Wittgensein's notion of family resemblance can serve as a corrective to the dominant approaches to the study of language.

Pär Segerdahl, Language Use, Ch.9, "Formal Pragmatics"

Language can be represented by a formal system, but that doesn't mean "that actual language use is based on tacit calculations" in the formal system. The meanings of the parts of a sentence play no role in the actual process of the determination of meaning.

Pär Segerdahl, Language Use, Ch.8, "Rationality as a Basis for Language Use"

Grice's cooperative maxims and conception of the rational norms governing conversation are meant to explain our ordinary practices of understanding sentences; but the maxims and conversational norms themselves depend on those very practices for their intelligibility. For example, the notion of cooperation requires some particular, concrete practice (like work in a repair shop) in order to have any content. It doesn't make sense to say that there is some general, practice-independent notion of cooperativeness that governs our understanding of speakers' utterances.

Pär Segerdahl, Language Use, Ch.6, "Literal Meaning"

Literal meaning is a function of use. Use is fundamental and cannot itself be explained. Semantic approaches to the explanation of meaning presuppose what they are trying to explain, because the paraphrases given by the semantic theory themselves, if they are to be meaningful, have their meanings determined by particular uses.

Pär Segerdahl, Language Use, Ch.3, "Context Dependence"

The meaning of a sentence is not determined by the meanings of its component words, it is determined by how it is used. Systemaic variations in the meanings of sentences should be explained in terms of variation in use, not in variation in their component parts. It is wrong to think that there are isolable linguistic items like words and sentences; a demonstrative like "this book", for example, includes the book as part of the symbol. The question of what must be added to a linguistic item, like "this book", to deliver a unique reference is confused. It is confused because it treats language as separated from the world.

Pär Segerdahl, Language Use, "Deixis" (Ch2)

An essential feature of the pragmatic approach to studying linguistic phenomena is the treatment of linguistic meaning as "general directions" (Strawson) for use. These general directions are supposed not to require any contextual features for their use (otherwise they wouldn't do any explanatory work[?]). But the general directions do require contextual features for their use, no less than words like "I" and "you". So giving the meaning of "I" as "the speaker of the context of utterance" is in fact no explanation at all. And it is only the conception of language as an autonomous system that makes the meaning of words like "I" and "you" seem problematic and in need of explanation in the first place.

Pär Segerdahl, Language Use, "Language and Context" (Ch1)

Pragmatics incorrectly assumes that only the meaning of special sentences (like those containing indexicals) must be explained in terms of the context of use. Instead, every sentence (for example, those including expressions like '9:45am') has a meaning only as a result of our linguistic and non-linguistic practices involving the sentence. Meaning is not primarily a property of linguistic expressions, but of our practices.

Pär Segerdahl, Language Use, "Introduction"

The "pragmatic approach" to language is an attempt to defend the autonomous "calculus conception" of language against apparent problems. The pragmatic approach attempts to handle those problems in the following ways: (1) Indexicality, (2) Speaker meaning; (3) Speech Acts; (4) Presupposition. The basic problem with the pragmatic approach is that it is circular--its application depends on those facts it attempts to explain.