These days, in the matter of climate change, simple epistemology has become a matter of dispute. Competing visions prefer appeals to emotion. After a plausible beginning some three decades ago, testable hypotheses concerning climate have faded into the background—eclipsed by an ever-ramifying and near-impenetrable tangle of acrimonious accusations, ad hominem arguments, well-poisoning, and appeals to authority.

The beginning of wisdom in such circumstances must be to sort the various assertions into two clusters: those that can now be settled to the point of moral certainty and those that remain within the realm of honest dispute. It is not necessary to immerse oneself in arcane scientific detail to do this. Observation and logic will suffice. The efforts will be aided by recently revealed facts, including the texts of internal communications of some of the principal actors.

Some of the frequently heard assertions may now be fairly judged as false to a moral certainty:

“The science is in, settled, or enjoys overwhelming consensus.” One needs to know nothing of the physics, chemistry, or biology of climate science to judge this frequently heard assertion as plainly false. Catastrophic change has believers and skeptics of equal eminence and probity. To claim consensus by excommunication from the lists of fair debate is despicable as well as logically untenable. More than that, science is never finally settled, and consensus is the stuff of politics.

“Those who disagree cannot be trusted because they have a vested interest in the outcome.” It cannot be clearer that there exists a moral hazard—a view shared by believers and skeptics. Contemporary scientific research is accompanied by massive sums of money, academic tenure, and, occasionally, access to fame and glory. Granting authorities, both private and public, have preferred outcomes in mind, and these are well understood by the grant applicants. The statement does not discriminate between believers and skeptics and is an empty assertion.

“Even if the current findings are uncertain, the application of a ‘precautionary principle’ requires that we act to avert catastrophe, just in case.” Again, the assertion is logically untenable, since it assumes what it purports to prove. Of course we need to be prudent, but we can’t know how to do that until we know the relative risks, benefits, and costs of alternative actions. This is precisely what the whole controversy over global warming is about. It might be that the expenditure of trillions of dollars in an attempt to alleviate a possibly trivial climate change would produce far greater woe.

These three propositions, still frequently proclaimed, serve only to distract and mislead. Reasonable debate should not involve their use.

As for those controversies that cannot be settled by the use of logic alone, they are of two types: questions of process and of content. The problems with the process of climate science begin with the corruption of the peer-review process.

Peer review of scientists’ work is necessary to the scientific enterprise. It requires open sharing of original data and recognition that colleagues, including hostile ones, may detect errors that confound our fondest hopes. Peer review is no guarantee of sound science, but it is one indispensable safeguard against avoidable error. The essential condition of peer review is that the peers not be deliberately selected by journal editors to be predisposed to agree with or condemn the work of others. This is a serious hazard, especially in disciplines that have only a few real experts all known to one another. It depends entirely on the integrity of the editors and the peer referees. In addition, the reviewers may sometimes be tempted to intimidate journal editors, threatening to withhold their future work. Thus the process is never entirely free of bias, human nature being what it is.

In the case of climate science, corruption of the peer-review process appears to have taken place. Communications among some of the principal investigators suggest a conspiracy to prevent the publication of work at variance to their own. In addition, they attempted to take action against editors and journals that published the work of their rivals.

Worse, these same investigators refused to disclose their original data and their methods of analysis, threatening to destroy data rather than comply with freedom-of-information demands, as required by law. This action constitutes scientific malfeasance of the gravest type. Alone it is sufficient to discredit their entire enterprise.

A second problem with the process is the corruption of the original data. At first it might seem that measuring temperatures around the world over many years would not be conceptually difficult, even if highly tedious. Alas, this is not so. In order to develop a comprehensive climate history, one would need to have solid data stretching for at least a thousand years. Since reliable thermometers have been available for little more than two centuries, it is necessary to make inferences from so-called proxy data.

Proxies may include measuring of tree rings, drilling of ice cores, measuring of mud sediments, or even use of historical anecdote, such as the freezing of rivers and the geographical distribution of crops. The accuracy of such proxies is always problematic. Even in recent times the collection of data is no trivial exercise. Where should sampling stations be located? How many? How often should temperatures be recorded? On land or ocean? In cities or countryside? No consensus on data-collection methods exists, and each investigator has his preferences, leading to data that are not always compatible.

Nevertheless, an estimate, however flawed, of the secular trend of temperatures for the past thousand years was necessary for any plausible projection of future trends, and for any estimate of the effects, if any, of human contribution to the variations thus agreed upon. Thence began a creative exercise in the “adjustment” or “correction” of raw temperature data. It is inevitable that such estimates contain a measure of subjective interpretation.

These interpretations must not have been made with complete confidence, since the methods of calculation have not been disclosed. Indeed, at some uncertain time, custodians of the original data appear to have deleted them, retaining only the . . . er . . . adjusted estimates.

So we will never know, with adequate confidence, what the temperature trends were thought to be by those who have been charged with custody of the many years of data on which, they insist, the future of humanity depends. Although there are four main foci of such data (the Climate Research Unit of the University of East Anglia, NASA, NOAA, and Darwin, Australia), they share some sources, remain unavailable to independent assessment, and show the same casual approach to integrity of the data. Requests for disclosure have been refused. This is a curious posture for publicly funded organizations.

As to the problems of the content of climate science, in spite of strident claims, none of the following questions appears to have been settled: Is present climate change, if any, outside historical parameters? Is there an ideal average global temperature? If so, what is it? Are changes on a positive feedback trend, and thus dangerous, or a negative feedback course, and thus self-correcting? What is the relative contribution of atmospheric gases, solar variance, ocean currents, and other as yet unknown factors? Is there an ideal level of atmospheric carbon dioxide? If so, what is it? To what extent does human activity affect the atmosphere?

The work done by climate scientists has not begun to answer these vital questions. The sources and methods are insufficient for the task. Data compromise, bickering, tribal loyalties, moral hazards, and perverse incentives all combine to form a cascade of error from which the truth cannot be rescued.

The course of this controversy begins to suggest the possibility that errors of method have not been entirely innocent. Questions of deliberate fraud and hoax remain unsettled. If they remain unaddressed, the popular impression of the integrity of science may be fatally compromised.

Nevertheless, these are important questions. It is plausible that human activity could be having some negative effects. We need to try to find this out. But it now seems likely that we’ll have to start over, returning the scientific enterprise to at least an adequate measure of objectivity. Perfection is never achieved, but we simply must do better than the current efforts. If there are serious dangers, let us study the costs, risks, and benefits of proposed interventions, a task that so far appears to have eluded us. Let us restructure the methods of investigation so as to minimize the distortions of groupthink. Let us start with the best data we can find and work toward a conclusion, rather than the reverse, as we appear to have been doing so far.

The burden of proof for destructive climate change firmly rests with those whose remedy requires an overturning of economic and political assumptions without precedent. We need to apply the best thinking of which we are capable. We haven’t done that so far.

In the postmodern dispensation that now beguiles us, this will be an uphill trudge. It is always more fun to damn the facts and embrace wishes. The great game of climate-change baseball is in the late innings, but Reality bats last.

William Anderson teaches at Harvard University.