There’s been a lot of rubbish written about “other ways of knowing”. So it’s quite refreshing to read Richard Carrier’s classification of methods of knowing. This is from his book Sense and Goodness Without God: A Defense of Metaphysical Naturalism. Well worthy reading by the way.
He starts by pointing out that no method of obtaining knowledge can produce absolute certainty. We can always be wrong, make mistakes. But we can list possible methods in order of reliability:
“What is rational is to assign degrees of conviction to degrees of certainty established by a tried-and-tested method. What is rational is reasonable certainty, not absolute certainty.”
“The methods of logic and mathematics are well-developed and provide the greatest certainty we have yet been able to find regarding anything, other than a present, uninterpreted experience. The next greatest certainty has been found in the application of scientific methods to empirical problems. In third place is our own daily experience, when interpreted with a logical or scientific mindset. Fourth is the application of critical-historical methods to claims about past events. Fifth is the application of the criteria of trust to the claims of experts. Sixth is the untested but logical application of inferential generalizations from incomplete facts–that is, plausible deductions. Such is the scale of methods that we have historically been able to discover and confirm as effective.”
“Experience shows that our degree of certainty will generally be weaker with regard to facts at each stage down this six-rung ladder, though within each category lies its own continuum of certainty and uncertainty, and the ladder itself is a continuum of precision and access to information: the more data we have to ground our conclusions, the farther up the ladder we find ourselves. Thus, mathematics is just perfected science; science, perfected experience; experience, perfected history; and history, perfected attention to experts; while plausible inference is what we are left with when we have none of those things.”
“Lacking any of the above approaches to the truth, we are faced with untrustworthy hearsay and pure speculation, where only the feeblest of certainty can ever be justified, if at all.”
Carrier writes that accurate methods of knowing have the properties of predictive success and convergent accumulation of consistent results. However, these should be evaluated intelligently. Even the best method may produce faulty knowledge if used incorrectly.
So how do the different methods rate?:
1: Logic and mathematics
Produces the broadest, most complete and most consistent results. The methods are relatively simple and they involve few, precisely defined, predictions which are easily validated (as Carrier says – “in the laboratory of the mind”).
Logical claims are about the meaning of concepts, not details, and this limits the applicability of the method. It is also easily (and often) manipulated. For example logical arguments are presented as arbitrary lists (proving only that the “logician” can count) or based on shonky premises – chosen to produce the desired answer. This often happens with people arguing for strong ideological prejudices.
I think this can be countered using careful validation by other logicians or mathematicians. Many of the conclusions can also be validated (or proved wrong) by application of empirically based methods of knowing.
2: Scientific method
This is actually a whole complex of empirical methods, and also includes logic and mathematics. I think it is important to see that the method cannot be reduced to a simple algorithm. In reality science can be quite messy and influenced by subjective desires. But it also includes processes to reduce the influence of subjectivism and test resulting conclusions.
The method is not as certain as the logical-mathematical method. It is also a complicated, expensive, difficult and often lengthy process. Requiring special care and extensive evaluation. “But,” Carrier points out, “when these standards are met, well and properly, our conclusions will be the most certain we can achieve about facts outside the human mind, correcting even our own errors in direct experience.”
We need to appreciate, though, scientific knowledge is relative, always open to change and improvement as we acquire more empirical and logical information. And often science needs to talk in terms of probabilities rather than absolutes. On the other hand we are often able to quantify the probabilities involved.
Ultimately all out knowledge comes from our personal life experiences. And we know that knowledge is largely correct because very different people agree on these conclusions.
But simple unexamined isolated experiences are not as trustworthy as many, well analysed and verified experiences. So we must accept that our knowledge based on personal experience is wrong if this is shown by science, logic or mathematics. The lesson here is that it is always best to examine out own experiences with logical reason and scientific honesty.
Richard Carrier points to the clear advantage of a personal philosophy of scientific naturalism. “For us, if we want greater certainty rather than less, the method of personal experience ought to be the simple practice of living a life of reason, applying scientific and logical principles whenever and wherever possible. This will ensure your life experience produces more reliable knowledge, and is more flexible (by being more open-minded and skeptical), and thus less challenged by the findings of science and logic.”
Because our evidence here is indirect, historical knowledge is less reliable than logic, scientific and personal experience. And methods available to verify or confirm this knowledge are also indirect and less secure.
However, critical historical analysis can avoid or limit some of these problems. It’s also important to realise that some of the criticisms of “historical science” are mistaken in that that knowledge can often be derived from several different lines of empirical evidence, often derived from current measurements, which converge on a conclusion.
5: Expert testimony
This is important because most people rely on this sort of knowledge for often very important decisions.
Expert testimony is essentially derivative of the other methods. For example scientific experts may derive their authority from actual involvement in the scientific, logical and mathematical methods.
Experts will clearly provide more reliable and trustworthy knowledge that non-experts. This places importance on criteria for determining the reliability of experts. Its worth “testing” them for reliability.
- Are their qualifications relevant to the questions at hand;
- Is their testimony confirmed by other reliable experts
- Is their evidence that the experts adhere to reliable methods of gaining their knowledge
- Do the experts make an effort to avoid or correct for their personal biases.
Clearly experts can be wrong and ideally their advice should be checked by other more reliable methods where possible. Their expertise counts for nothing if their advice conflicts with knowledge obtained logically, mathematically and scientifically.
These are important qualifications for the person in the street who often relies on expert testimony for input to their own important decisions. Just consider, for example, the political importance of expert testimony when considering climate change and political decisions arising from it. Unfortunately, many people “choose” their expert using confirmation bias rather than objective assessment. The advice from Richard Carrier on the personal advantage of scientific naturalism (see 3: Experience above) is relevant here.
A claimed area of expertise may be inappropriate to the question at hand. For example, militant theists will often argue that comments, articles and books written by scientists, philosophers and others questioning existence of gods are irrelevant becuase these people are not theologians. As Richard Carrier points out “a theologian may be an expert on theology, but that only means he has a genuine experts in concepts of theology, not that he is an expert on factual questions like whether a god exists or whether Catholicism is the One true religion. No one can be an expert on these questions becuase no one has any real evidence for them, at least evidence properly produced by one or more of the superior methods above. A theologian can hardly claim any more experience with an actual god than we can.”
And we need to recognise that in some areas “like theology we find very little agreement among qualified experts, and a vast influence of ideological bias that is rarely placed under any objective control.”
6: Plausible inference
It is reasonable to trust plausible inference and inferential generalisations if well argued. But we shouldn’t give these more credence than the more reliable methods of knowing.
I believe this method has an important role in science and should not be rejected just because the evidence is incomplete or missing. Speculation and wild ideas are an important source of creativity and of hypotheses for testing.
In fact some ideas or hypotheses based on plausible inference may have useful explanatory power and be useful where validation is not yet possible. Consider “String Theory” and the “Multiverse” ideas.
However, we should expect that a proportion of ideas based on plausible inference will fail when tested scientifically. This is a salutary lesson all good scientists learn early in their career.
7: Pure faith
These are beliefs based solely on tradition, hearsay, mere speculation, desires and wishes. Beliefs in ungrounded assertions.
We know from experience that such beliefs usually turn out to be false. Just consider all those legends, traditional myths and superstitions which have been shown wrong throughout history. Yet the method of pure faith transmits beliefs without any regard to their truth. Faith conveys false beliefs just as well as it does true ones.
So the probability of faith-based beliefs being reliable must be low. Carrier writes: “blind faith is inherently self-defeating. The number of false beliefs always vastly outnumbers the true. It follows that any arbitrary method of selection will be maximally successful at selecting false beliefs.”
Some sense at last!