Although I can’t see any reason to believe that our thoughts extend beyond our minds, or require an external world to have content or exist, it is an undeniable fact that some things are external to us. Obviously physical objects and events are one such thing. More interesting however is that the truth of statements is also external to us.
For example the truth of the statement “the sun is the center of the solar system” is independent of what people think its truth is. This is to say the truth of the statement depends only on the nature of the sun and the solar system, and not on human thoughts. (Of course an obvious exception to this rule is any statement that refers directly to human thoughts.) Another point to bring up is that I take it as a given that the statement is well defined (that each of its words has a well understood meaning). This is not to say that the truth of a statement is dependant on the meaning of words, simply that because words are one of the few methods we have for communication a failure to understand what the words mean will result in a failure to understand the statement. Relativists of course argue that because people understand the meaning of words in different ways there is really no truth that is independent of the individual, but I say that in a case where people understand the same word in different ways they are really just considering different statements, and the truth of the statements they each consider is external to them.
Unfortunately this limited version of externalism can lead some people to believe that externalism is true with respect to thoughts as well. This happens for two reasons. One possibility is that people accept Frege’s definition of meaning, namely that meaning is defined by how a word contributes to the truth of statements. Accepting this they then assume that since meaning is directly connected to the external world, and because meaning is a key component of thought, then thought must also be directly connected to the external world. The mistake being made here is a failure to distinguish between two kinds of meaning, the objective meaning and the personal meaning of a word. Although the objective meaning does depend on the external world the personal meaning may or may not, and thus it is bad reasoning to assume that because the objective meaning depends directly on the external world that the personal meaning, the kind of meaning that is a component of thought, must do so as well.
The second mistake people are sometimes led to make is to assume that because our thoughts often take the form of statements that thoughts must, like statements, depend on the external world. The mistake being made here is that it is not the statement that depends on the external world, but the truth of the statement. I accept that the truth of certain thoughts might also be said to depend on the truth of the external world, but since there is no indication that our minds have direct access to external truth of thoughts (except through perception) there is no reason to say that our thoughts are dependant on their objective truth or falsehood. (This of course also explains how we can be in error and not know it, as there would be something strange in saying that our thoughts are dependant on external truth and yet that we are unable to use that connection to our advantage.)
Of course just because our thoughts aren’t dependant on the objective truth doesn’t mean we can’t have access to such truth. A trivial case of course is our ability to know the truth or falsehood of statements that are tautologies or contradictions. However most interesting statements are neither true nor false necessarily. We discover the truth of such statements through perception and experience. We judge that probability of a statement is greater given our perception of certain evidence (even if this “evidence” involves direct perception of the contents of the statement, since there is a small possibility that our perceptions could be mistaken). It is true that given a single perception we are unlikely to be absolutely sure that the statement is true or false, but given repeated observations the probability of the statement being true generally approaches one or zero. (Because if Pr(s | m) != Pr(s) & Pr(s | n) != Pr(s) then Pr(s | m & n) != Pr(s | m)) Of course no matter how many observations you make nothing can be absolutely certain, but we can be certain enough for all practical purposes, enough to say that we know the truth of various statements.