I very much prefer the 'bildung' school of philosophy talks, which are meaningful both in their relation to other things and in themselves, but the trend is definitely towards those that only answer to the first requirement. I remember a talk one evening at a large state university that solely consisted of 'calling out' to the people in the audience who had done work in the field and offering a thought or two on each line of work, and leaving the interaction to the conversations afterward. (I headed out for a solo coffee.)
When a talk is only meaningful in relation to other things, like a jazz line that requires some knowledge of the original melody to have any meaning other than patterned exertion, it just means that the world of the event is larger than the room. An obscure point on epistemology in relation to someone else's article, even if it does make a new theory of mind snap into place for the first time, does require a distributed cognition of sorts. And the myth is that it's all one big cognition, slowly reaching its perfection (and a dangerous objective correlative is emerging in the mechanical world). In fact, these integrated speakers are simply responding to the events in the near space, without really advancing the cause of any larger project.
Heidegger's definition of a discipline, as opposed to an area (I think), was the creation of a reticulated shared field of reference associated with the world at the point at which it claimed a competency. A theatre faculty, for example, would be doing meaningful work so long as they all knew what each other meant, and shared the sense of the relations between the terms that they used, and these terms were associated with objective phenomenae of the art at certain points. The difficulty, though, is that most faculties make a much broader claim than that, and assert that they are connected to the whole, both in the particular and in the meaning that they create, but for that, you'd need to either have a very good philosophical understanding of the relation of the art form to the world, or a complete lack of a philosophical understanding of the relation of the art form to the world. The result of any discipline is either the transformation in the thing that it is studying (politicians learn history, mice master the maze) or a reliable map of the territory, although labelled in an imaginary language.
To connect the earlier thought, their relation to the melody might be real, but since the one outside the field doesn't hear the music as they do, it just seems like patterned exertion. (Many lit-crit/junket conferences consist entirely of these trading-eights.) But the point is that these patterns should be reliably associated with objective elements of the studied practice, either mapping it or transforming it.
Science has displaced personal understanding -- yes, the language of a given field, so long as it is mathematical (Kant's criterion) or experiential, is the mensura, but only according to its own logic. Humans still govern all talk of the totality of things.
Say someone totters into the agora one day and shouts: "It's all water! Everything is water." A short time later, someone thinks to disagree, and suggests that it's all fire. Now, as the discussion begins, no one is under the illusion that they are going to be able to discover the water or fire in things. What makes it meaningful is that a human, at the most perceptive condition they could summon, and in the public square, spoke their best truth, and there was some truth in what he said. When we respond to the idea, we respond to the soul who fashioned the proposition sive necessitatis.
And even the most dry talk, if you follow the cloud of distributed cognition, ends in a sacramental profession along these lines. If you were to spend your life arguing with the propositions generated and advanced by a computer (without considering them to be derivative of words spoken in the air), you'd rightly be called a computer repairman. The meaningfulness of the proposition isn't in the the object, but the res. The human, in relation to the situation, who fashioned a thought.
So I'm partial to the educative approach (interesting back-and-forth in the TLS letters recently about educare/educere) in which you give at least some sense of the argument from first terms, largely because this tends to put the responsibility on the speaker to say something themselves. Much as I think the very old practice of judges all writing separate opinions and letting the bar figure out the law of the decision was a much better scheme. The human voice, and the apodictic. And it's best if both the voice and the truths are in the same room