Why Derrida? Why Grammatology?

I've been rereading Grammatology over this summer and it's left me with some need to revisit 1) why Derrida has interested me, and 2) why Of Grammatology has interested critics in the US.
If you'll permit the first autobiographical reflection, let me begin by saying I never really understood what he was saying until about a year ago, and I only did so because I just kept reading as much of him as I could. So the question really is why I continued to read him when I didn't really understand what he was getting at?
I was introduced to him in a class on theory--like many students, I think. I was initially very resistant to things that he was saying, being more Kantian and Hegelian in spirit, and more interested  generally in philosophy of mind. But what turned me on about him was spending a lot of time with a few essays of his--and writing papers on them. What was so great was that he was extremely rewarding if you read him closely and tried to piece together what he was saying. Some might call this his "play on words" or whatever, but what was really interesting to me was the high rationalism at work in the construction of his sentences: like any good novel, you can really get a lot by working through them. This made me less resistant.
As I took more and more classes in philosophy, which I was pursuing at the same time, I began to see that this sentence-level aspect of the way Derrida wrote was neat, but ultimately less interesting to me than the way he worked with the philosophical tradition. As I personally became more interested in phenomenology as I was pursuing philosophy of mind, I found he plugged himself quite thoroughly into that sphere. But this did not make me reduce him to this tradition--as some are too quick to do. I gathered that there was a particular intellectual field that was very diverse and yet very coherent that was orienting his discourse, of which phenomenology was a major part. And I began to piece this field together. In short, the tradition of French philosophy from Canguilhem on became an interest to me.
But at the same time, I just kept reading Derrida. And what was interesting to me then was his way of reconstructing other works: in short, I found that Derrida was actually an amazing teacher in his written work. He would have to reconstruct the entire logic of a discourse, and what this produced was actually a really neat way of introduction into a particular way of thinking about whatever he was talking about. Most might presume that the reconstruction would have to be skewed or distorted, but usually Derrida is also reconstructing how the field he is working in is thinking about the particular issues. Side comments and little aphoristic remarks are what allow you to grasp this. Indeed, they are unintelligible, but what was striking to me was not that they are unintelligible because they play with words or whatever: they are unintelligible in their content, because they presuppose an entire set of arguments and indeed a point of view in looking at whatever text you are considering. In short, I liked his patience--not because it produced rigor, but because it was symptomatic of that field in which he worked. A good example of what I mean is his discussion of Freud and the mystic writing pad in Writing and Difference. There he tackles Freud's 1895 Project for a Scientific Psychology, which many people in France were rediscovering at the time. He reconstructs the logic of this Project, and in doing so allows you to see what people are talking about then. But at the same time, he gives you a take on it--and all of this happens before he really begins to fiddle with what is going on. These little reconstructions and set ups--the way he poses the problems, were really neat to me, and what kept me going even though I didn't understand where the hell he was going with them.
Thus, a lot of emphasis upon the key words Derrida gives critical theory (trace, etc.) was completely lost for me, and mostly willingly--this was a healthy sort of ascesis that actually helped me not to use much of what Derrida said in my work on literature. It also kept me from immediately using him in my work in philosophy: my concern with Heidegger and Merleau-Ponty and others seemed to have its own impetus. And yet, as this reconstruction of mine makes clear, one can see that both these interests--in literature and philosophy--often take their cue from Derrida, or orient themselves around trying to figure out what he was about. At a certain time, then, it all started to pay off in the form of more and more interconnections, and I really got a sense of where he was coming from--in other words, the field was reconstructed. What was made clear, however, was that Derrida was more at home as a thinker within the tradition of philosophy than in critical theory and the study of literature--I found most people who talked about him there were really too far away from the study of literature at the same time as they were horrible at philosophy. However, his spirit was more within these other spheres, precisely because they addressed questions (gender, for example) that were escaping philosophy (although I know now that in certain spheres--like ethics--they weren't and aren't).
If I had to do it all over again, I think a quicker way of understanding him would be just to go to France and work in their philosophy departments. But I don't like the way they work as much as the Americans, as odd as we are, and so I think I would have lost a lot. Quicker wouldn't necessarily be better, I guess.
But all this brings me back to Grammatology--my second point. I read this first, I think, after some of the short essays that first introduced me to Derrida. And reading it then, sort of passively looking for those points of orientation I spoke about above rather than for the definitions of things like trace and differance, on the one hand, or the overall argument, on the other--reading it then and in this way was something that actually made the book much more coherent to me, I find.
Reading it now, the thing seems pretty crazy. Huge, sprawling, disconnected, extremely radical, the work is much less clear about the key concepts of differance and trace than Speech and Phenomena, and at the same time is so massive in its force and its demands upon thinking generally that it isn't clear where its argument most hits at other than the entire philosophical tradition--something that is done more economically, I think, in Dissemination and certain essays in Margins of Philosophy (e.g. "The Pit and the Pyramid," "Ousia and Grammê," or "The Ends of Man"--take your pick). So it isn't clear to me why people turned to it and still turn to it to understand Derrida.
And yet, I somewhat understand it if I focus on the analysis of Rousseau, which is quite simply just brilliant. However, this is so massive and, again, sprawling, that I don't know if it has the same effect now as it would when it first came out--that is, when there was a certain consolidated understanding of Rousseau as, well, merely contradictory (not focused upon and determined by something like a supplement). Better now to tell someone to pick up Speech and Phenomena, again if they want to begin to see what Derrida is about. In other words, Grammatology is all about extending to the utmost what is accomplished in the analysis of Husserl, because it actually grasps the hugeness of the consequences of that analysis in Speech and Phenomena.
What are these consequences? Simply put, that a phenomenology of text would be impossible. And, in fact, more than impossible: it would require, if one did not just back down in front of it, a total revolution of phenomenology, where the concern with essence upon which it is founded has to be completely evacuated. Phenomenology would have to proceed without phenomenology, then. Text becomes a privileged phenomenon, then, within the tradition, because it disrupts it completely, as well (and here is where Rousseau comes in) as anywhere there is a concern with essence.
In a way, though, you need to understand the basic thesis before you can grasp these consequences extended throughout Rousseau and (in the beginning of the book) Levi-Strauss. And--here is what I really think--the first part of the book does not totally introduce you to them as clearly as other parts of Derrida, notably Speech and Phenomena.
That said, it is clear that what really is great about this book is how you see the "method" of analysis take shape--and I think this is why it was and is so popular as a sort of introduction or primer. What is also remarkable is the repetitiveness of the analysis: like Dissemination, what we get is a focus upon the logic of the supplement at every moment it seems to come up in the discourse. And yet, unlike Dissemination, we get more of a setup that inserts this work of focusing within the current discourse: the long discussion of the economy of pity, etc. with Starobinski et. al. is very odd, actually, because it doesn't clearly do what it seems it wants to do. This is a double task: 1) reconstruct Rousseau's discourse and 2) reconstruct our way of reacting to this discourse as indicative of our current (that is, at the time of Grammatology) way of looking at what threatens to revolutionize our philosophical moment. If one focuses on the effort of reading to do this, one really sort of mistakes what Derrida is doing here--quite simply because it is too much setup and makes the real reading have to account for too much. In other words, this is too external to the task.
Where Derrida is really reading is when he is describing the supplement and showing how it always can be explained two ways. In other words, the whole effort of peeling back what Rousseau describes--which is the supplement--from what he declares--which is either that the supplement is bad or good in a particular case. Here the "method" is happening, and one sees how you have to go about it. You have to try and enter the protocols of the text, as Spivak often tells her students. This sort of vague phrase means really that you have to reconstruct a logic that the supplement, which functions as the center of this logic, its anchor, will have to disrupt. Getting a handle on this disruption means moving back and forth between logics, trying to make the structure of the disruption--differance--appear in how it leaves a trace in all those logics--or rather how all those logics become not logics but traces. In short, you do not just point out where differance is operative, where what Rousseau declares is different from what he describes. This is to say Rousseau thinks something "without thinking it." Rather, you show how the description is coextensive with the declaration--that is, you show that differance is coextensive with the articulation of the discourse itself. This is what Derrida means in the following:

It does not suffice to say that Rousseau thinks the supplement without thinking it, that he does not match his saying and his meaning, his descriptions and his declarations. One must still organize this separation and this contradiction.
-Of Grammatology, 245.

And this organization in fact exceeds what it organizes, as Derrida goes on to say. In the end, reading Grammatology, you really begin to see how this work of analysis functions--or at least what it would require. I personally think Dissemination ("Plato's Pharmacy") shows it more purely or at least more cleanly, but as it is, Grammatology remains a great example of this method or work--which is of course not a method and, as a working, an unworking.
But back to Rousseau: why this work is so amazing is not necessarily because it can account for things we that merely seem oddly contradictory in Rousseau ("I am the least vain of anyone," he says somewhere in the first book of the Confessions and remains for me the greatest, most hilarious example) as I might have suggested a bit earlier, but because this account proceeds in the way it does. Quite simply, it shows that Rousseau is a modern thinker of phusis, growth and generation. And at the same time it shows that this thought of growth is not possible for the moderns--a thought that in fact makes it possible that the ancients never really thought it either. Nature, that amazing force in Rousseau, is so conflicted that this becomes absolutely clear: however, its remains conflicted not because it is pure, but because it always has to articulate itself. In other words, because it is impossible, and impossible in such a way that it is the only thing that makes the discourse of Rousseau in its totality possible. Hegel, Marx, Freud, whoever--these thinkers might have very crucial parts of their discourse governed in this way by impossibility, but no one has it so totally and so crucially articulated it as Rousseau.
I could go on, but all I wanted to do was bring out both my reasons for reading Derrida and my reasons from shying away from various aspects of Grammatology as an introduction to Derrida--with the requisite highlighting of what is indeed important about it, which I think is not what we usually hear. We usually hear all this stuff about text and writing, but this is only the first part of the work (which I think many simply are content with reading, after which they put the book down). But all this is outlined in Speech and Phenomena and Writing and Difference. What is unique and important about Grammatology is that it works out the way that this problem of writing and text diffuses itself or allows discourses to be determined by it. Thus, it is appropriately wild and sprawling and odd--it has to exhaustively show discourse as trace. But this I think needs to be now seen as its virtue more than the simple thesis about grammatology as a positive science. If one wants this, one should look at a more pure, more clean analysis of Derrida's.
In other words, the sprawling, exhaustive aspect of the analysis here (the fact that it accounts for so much in Rousseau, not as much the aspect of the analysis that deals with inserting it into a current debate on Rousseau) has to be looked at as its odd core--its dispersion as its coherence, in a nutshell--because it is the most interesting working out of a way to talk about that impossibility: it is the work of analysis and reading when it is directed at what is impossible to analyze and read. In short, I think looking at it as less a finished work than a work in progress, a work of reading in progress, is necessary to appreciate it and indeed appreciate its genius. Looking for ideas like trace and differance needs to be done elsewhere. The analysis of Levi-Strauss should be looked at more in this manner too--and less as an attack or a condemnation of modern Rousseauism. But I might speak about this later--I've shied away from Levi-Strauss here so as to keep things simple and reconstruct my main thoughts about the book and my relationship to it.