Phil wonders about the standardisation (or not) of language. I was going to post this as a comment, but it got big enough to warrant its own blog reply...
"Search engines are a powerful incentive to keep to some kind of standards."
An interesting technical reason (and the same could be said - moreso? - of tagging), but I come at this from a slightly different point of view - does the ability to spell indicate some level of intelligence and if so, is that a cultural bias or a more instinctual, "group seeking" one? I started to think about this before but never came back to it.
I also find that I make the mistake between "they're", "their" and "there" far more than I do with "its/it's" or even "to/too" - maybe theyre's less nit-picking over it, so you look out for it less...
Anyway, my question is - are we right to use spelling and grammar as an indicator of "intellectual validity"? And/or do other factors come into play depending on the arena in which the content is?
For instance, spelling mistakes in an academic paper are considered far more serious than spelling mistakes in a blog entry, due to the air of quality and checking associated with the former. That air of quality is similarly applied to the concepts in the academic paper, which translates into a reputation for the content, regardless of what's actually written. Therefore, their are inherent associations we make between mistakes that we see, and the quality of content that we take in, but those associations are largely a result of the mechanisms that have built up around the publishing methods employed.
That's one aspect, as I see it. The second aspect (which I think I touch on in the "article" linked to above) is that there may also be a more "individual" (i.e. not bound to the infrastructural reputation mechanisms) connection between SPG and intellect that we, as native speakers of some language, make without thinking. That is, understanding and comprehension of a language indicates a certain level of thought, which translates (correctly or not) into a similar level of being able to pick up non-linguistic concepts, etc.
This would be fair if, say, the "model" that underlies any language is similar in comprehension to models for anything else. i.e. Language has rules, just as everything else has rules. Understanding those rules may or may not be the same for both linguistics and non-linguistic subjects, but if you can "get" the rules to the former, their's more chance that you "get" those for the latter too.
Thirdly, language is more than just an agreement. It can also be considered, I think, as an esotericism - a platform that has many non-obvious intricacies that require a thorough grounding in the language to understand - things like allusions, puns, cultural knowledge, and so on. I have noticed over the past year, for instance, that talking to foreigners is profoundly different to talking to English people - that level of tacit understanding - "unspoken implications" - achievable in any language speaks volumes. Unfortunately (but necessarily) this level often gets stripped out under "translation" (global English vs "local" English, say), or when using transfer means such as e-mail. While not completely confined to spelling errors and such like, the same idea that "linguistic knowledge == ability to understand tacit inferences" may prevail.
A last point, related to the first 2 above, is that (as Phil mentions) mistakes indicate "rushed" content. Perhaps the absence of mistakes indicates an ability to check over what you've written, which further indicates a more thoughtful approach to the content - if you've gone back and checked the lexical side, maybe you're brain has had time to process things semantically/logically as well.
In that sense, then, the gatekeeper is in all of us...