When a person speaks he or she enunciates words one after the other, producing a temporal sequence of words. Normally these words are used not mentioned. It is natural to assume that nothing meta-linguistic is going on: one word is used, then another, then another, until the utterance is complete. A theorist may mention the words uttered by the speaker, but normally the speaker doesn’t. Speech normally consists of sequenced word use, with no mentioning in the vicinity.
But is this right? Isn’t it true to say that the speaker combines words to form phrases and sentences? There is an operation called “combination” that the speaker (or his brain) applies to words. This operation permits a synthesis to be formed—there is an act of synthesizing words. But combination and synthesis are precisely operations on words. They take words as arguments and produce wholes composed of words as values. These operations are psychological in nature—indeed, they are intentional. The speaker intends to combine words into a sentence, or to produce a particular synthesis of words. The operations thus incorporate reference to words: “Combine ‘John’ with ‘ran’ to get ‘John ran’”. They are meta-linguistic; they involve the mention of words. But then speech inherently involves mention as well as use: the words uttered are used and mentioned simultaneously. We use words in grammatical combinations, and those combinations are the result of meta-linguistic acts or operations. The same goes for writing, perhaps even more clearly: writing a sentence is performing a combinatory operation on words. We use words in sentences as we write, but we do so by representing (denoting) them in acts of mental combination. When I wrote that last sentence I intentionally selected certain words and combined them in a certain way, referring to them mentally. Words are thus objects of intentional mental acts in both speech and writing: they are referred to as well as referring. The hearer or reader must apprehend words, taking them as objects of intentionality, but the speaker or writer is no different—he too must take words as intentional objects. The mind must mention the words it uses.
We can think of the speaker as following instructions (rules) for the production of sentences, and these instructions refer to words. That is essentially what a grammar is. So there is mention wherever there is use; use depends upon mention. Logicians and linguists talk about the concatenation operator, which connects quoted expressions–as in “John”^”ran”, to be read “the word ‘John’ concatenated with the word ‘ran’”. Grammatical rules can be written by invoking the concatenation operator. If we think of the act of linguistic combination or synthesis as (rule-governed) concatenation, we can say that speakers perform an act of concatenation on words whenever they utter a sentence—they in effect join one quoted expression with another. They are quoting themselves as they speak—mentioning what they are using. There is a double act of reference going on: “John” is being used to refer to John, but it is also being referred to itself in the act of concatenation. The total speech act consists of using words and also mentioning them. This is because speech is not just uttering words in temporal sequence but also selecting and combining them according to rules. In other words, there is a plan behind the utterance and this plan involves arranging words in a certain order. You choose words and you choose the order in which to arrange them: both of these are meta-linguistic intentional acts. Speech is therefore always about words as well as about things. Speech is always conscious of itself as speech. 
 Or unconscious of itself: we are not generally conscious of the mental operations that generate speech. But the operations necessarily take words as objects, so the mind-brain mentions them—the internal computations are defined over words. Of course, sometimes we quite self-consciously select what words to utter and we assemble them with care: then we are clearly consciously thinking of words as a preliminary to using them. In speech we focus on referring to things in the world by using words, but we also direct our minds to the words we use.