MANCHESTER, N.H. -- Preparing to come to Manchester last week, I dipped into the news clips of Iowa caucuses and New Hampshire primaries past and there they were -- the familiars: not just a number of the same candidates (Bush, Dole, Hart) but also most of the same patterns. The one that interests me most involves our seemingly infinite capacity to (1) chronicle and analyze these things till we could scream and (2) nonetheless again and again end up being taken by surprise. Why is this?

My tentative answer is that this does not occur despite all our hovering, meticulous attention, but more probably because of it. I think we are being drowned in a new kind of abstract analysis, political war-gaming when you get right down to it, that pretends to a realism and authority it does not have. And I think almost all of us, especially the candidates, have set too much store by these airy calculations.

First, consider the recurrent pattern of surprise. In the cold light of 1988, 1980 is particularly instructive: Bush pulls the inevitable Iowa upset; overnight he is retroactively explained and romanticized to the point of beatification; his every prior act, quirk and trait is now seen as an obvious component of his victory. At the same time it is being explained by the campaign advisers and gurus we consult why Ronald Reagan's defeat was inevitable, why he was never a credible, let alone invincible, front runner to begin with. Then Reagan trounces Bush in New Hampshire, etc., etc. Four years later something akin to this happens. You get the "Hart surprise" (not an outright win) in Iowa, enough for Hart to get the sudden glorification. But unlike the 1980 George Bush, Hart goes on actually to defeat the front runner Mondale in New Hampshire, who, still staggering around the ring, manages ultimately to prevail over Hart. As I write, the comparable stage of this by now set-piece battle between Bush and Dole has yet to occur, so I don't know which of these scripts it will more closely resemble. But what has already been demonstrated in Iowa is that, once again, we have been smothered in information, down to the pickiest detail, and that we still never seem to know what to expect.

Some overdue qualification here: there are certainly a number of brilliant political analysts around, both within the press and among the newly expanded priesthood of full-time, professional political consultants. What I am talking about is that sodden and unfortunately all too influential mass in the middle, the ones who seem to get candidates to do truly senseless, stupid things and somehow, miraculously, to get the rest of us to accept as gospel things that in our gut we just know aren't true -- things that don't feel right, don't comport with what we thought we saw looking at the same debate or speech or sequence of events.

"Well," we say, "they say Throgmorton won that debate. He looked like a perfect idiot to me. And awfully sneaky. But they must know." I think some among us so thoroughly internalize and incorporate this alien but intimidating "wisdom" that we may actually tell polltakers and one another that that's what we think, too -- but we don't. It is in that invisible crevice of distance between what we think and what we think we are supposed to think or have been told we actually do -- multiplied more or less by the voting-age population -- that the recurrent surprises are created.

The problem is an excess of clever but shallow imagining. It leaves out the way real people behave. In fact, the burgeoning new guruhood may be compared to the Kremlinologists of an earlier day and, in their wake, the fraternity of defense intellectuals. These, too, had their brilliant ones. But these, too, had their mass of initiates into the arcane lore and technique of their deliberations, so often conceptually logical and pragmatically -- humanly -- wrong. The first bunch were forever explaining to us, without a trace of shame, why the Soviet rulers whom they had so self-certainly pronounced in charge six months before had just been pitched out a window. The more bloodless among the defense intellectuals were forever describing nuclear-warfare scenarios that took account of everything except what would actually happen. God knows, we in journalism (and I freely admit to a mile-long list of such misjudgments) get people and probable outcomes wrong. And all too often, in addition to our own plain errors of analysis, we slide into the political priesthood's traps, so that we end up making not only our own mistakes but also theirs.

The point is that we are forever buying into interpretations that run counter to our own instincts and perception. I kept being told Bush did himself a world of good snapping at Dan Rather; in theory he did, a big "anti-wimp" win. But he didn't look good doing it. He looked all wrong; he doesn't do that kind of fighting well. Let a candidate turn up in a parka and a seed-and-feed cap instead of his usual garments. Let another announce that he has just acquired passion. The wise ones tell us the "new" position may win in the forthcoming contest. Or they tell us it will backfire, as in the insight being put out a few weeks back that Bush, by getting Dole to confront him over the Bush campaign's assault on him and his wife, had made Dole show himself as out of control and thus had lured him into behaving in a way to threaten his own Iowa lead. That sort of narrative is, it seems to me, almost always, the most unconvincing kind of plot-writing.

I don't think we ever take sufficient account of the possibility that the prospective audience for all this will not be moved, that they will not behave in the pat, slick way they are supposed to. The former Democratic chairman, Robert Strauss, says this kind of invention always puts him in mind of that great story about the dog-food company that spent millions on marketing strategy -- a host of comprehensive studies and spectacular ad campaigns to overwhelm the competition -- and, failing to sell any more dog food as a result, got the explanation why from a humble salesman in the field. "The dogs don't like it," he said.

I realize I am coming out for some primitive, untutored, unspecialized, perhaps even know-nothing version of political spectatorship. For all I know it could be anarchy. The premise, the discipline, is simple. No more accepting, and eventually even espousing as one's own, that preference or calculation or reaction that you know you do not share, that strikes you as baloney and which you fully intend to go against in the privacy of the polling place.

1988, Newsweek, Inc. Reprinted by permission; all rights reserved.

Why do we allow the gurus to get us to accept as gospel things that in our gut we just know aren't true?