Formal and informal discourse markers
There seem to be so many discourse markers used in spoken English that it is not hard to pluck a few from the linguistic ether to discuss, so I will select I mean, you know, oh and that discourse marker par excellence, okay-- four terms that rarely make an appearance in formal written contexts. Some of their formal written equivalents might include that is or i.e. (for I mean), in some contexts as noted or as is generally known (for you know), and the obvious alternatives yes or indeed (for okay), but probably only for a small minority of its spoken occurrences. However, I intentionally made this task harder that it needed to be by selecting discourse markers that frequently act as semantically vacuous fillers that lack formal written equivalents. I hoped to make the point that because speech and writing have very different communicative purposes, there will not necessarily be a formal written equivalent (i.e., "words or expressions which mark the same logical relation") for every discourse marker used in speech because not all discourse markers indicate logical relations. In speech, we often need to hedge, backchannel, reformulate and change topics on the fly, whereas writing is usually a much more planned and deliberate endeavor than speech.
Observations on so
The word so has a long and venerable history in the English language. As Germanic as Germanic gets, it started to wend its long and tortuous course through the history of the language as Old English swā at a time when its Gothic sister swa was alive and well, and similar-sounding words were flying from the mouths of Germanic tribesman all across northern Europe. In fact, the ultimate ancestor of Modern English so is probably at least six thousand years old, as its Indo-European cognates si, "if" (Latin) and ὡς, "as, thus" (Greek) attest. And oh how the Anglo-Saxons loved their swās--so much so, in fact, that they often uttered the word twice in a row, as in the phrase swā swā, which most famously appears in the line geweorþe ðin willa on eorðan swa swa on heofenum ("Thy will be done on earth as it is in heaven"), but interestingly did not mean so-so (as in "mediocre"). The word was so versatile in Old English that it could even be used as a pronoun in addition to its modern function as an adverb and conjunction (see Old English entry at http://en.wiktionary.org/wiki/swa#Etymology_2) and analysis of Old English texts reveals it to be among the most commonly-used words in that ancient language.
So, I must admit that I am not particularly nonplussed by the term's frequent appearance in Modern English or news articles alerting us to some sort of major shift in the usage of discourse markers in English. While Giridharadas states that "[w]hat is new is its status as the favored introduction to thoughts, its encroachment on the territory of 'well,' 'oh,' 'um' and their ilk," I have to wonder whether what is actually new is the word's increasing frequency or just the fact that we are starting to pay more attention to these sorts of things. The point I am trying to make is that the word so has always been with us, and even a thousand years ago it was among the commonest of words. So, even if this grand old word has started to take over ground once held by well or oh or um, as Giridharadas avers, our trusty friend so already started out with its fair share of discursive territory.