One of my student summer jobs was as field assistant for a geologist mapping the nickel-belt region in northern Canada for the provincial geological survey. For a couple days, we were joined by another geologist and I listened as the two of them discussed what might have happened to produce some of the rock structures we observed. They did not always agree. From the same observations, they inferred two different conclusions. They could not re-wind the clock and re-run the experiment. There were far too many possible variables, involving thousands of square kilometers of geological activity, to obtain the observed results. Instead, they could only draw conclusions based on inferences.
Inferring conclusions that go further than what we can experimentally reproduce is a very large part of modern science. This category of science can be called inferential science. We begin with observations or other experimental results to infer a conclusion that, itself, we are not in a position to experimentally prove. We can only say that, given the data, there is good reason to think it might be right.
The most common type of inference used in modern science is inductive. This is where the probability of the conclusion, given the data, is high enough to warrant the inference. It is this area of science where tension between science and faith often arises, not because of actual experimental results, but because of inductive conclusions that we are not in a position to reproduce.
So What Can Go Wrong?
As a scientist, I am often dismayed at the naïve faith in science that I see in the general public, including Christian leaders, in their tacit assumption that the interpretations and inferences of science are the final arbiter of how we must understand the Bible. They assume that if there is a tension, it is the Bible that must be on the defensive. In reality, science is no more immune from the frailties of human nature than someone’s interpretations of Scripture. For the following reasons, the inferences of science must be examined just as critically as any other set of beliefs.
A. Lack of Accountability
As I wrote in an earlier post, “perverse incentives” such as competition for funding, academic prestige, and the pressure to publish have resulted in a reproducibility crisis in science, especially in the biological sciences where the majority of peer-reviewed papers cannot be reproduced. If this is the case in experimental science, where methods and results are published so that others can try to reproduce them, resulting in a very high level of possible accountability, what are the implications for inferential conclusions that cannot be reproduced but the same “perverse incentives” still operate? In some areas, such as forensic science, the inductive conclusions are often tested in court, so accountability is quite high. In many other areas, the data and experimental support are so strong that the conclusions are very reliable. There are a few areas, however, where there is much greater freedom to get creative, knowing that there will be little in the way of accountability. For example, as biologist Austin Hughes wrote:
In recent years the literature of evolutionary biology has been glutted with extravagant claims of positive selection on the basis of computational analyses alone … This vast outpouring of pseudo-Darwinian hype has been genuinely harmful to the credibility of evolutionary biology as a science.
The bottom line is that we should be very, very cautious about embracing inductive conclusions that we are not in a position to hold accountable by experimental reproduction, especially narratives, even sophisticated computational ones, about the origin of life and the large-scale evolution of organisms.
B. Inductive Leaps on the Basis of Miniscule of Non-Existent Probabilities
The rational justification for an inductive conclusion is the likelihood it is true, given the data. It is often the case in forensic science that the evidence is so strong, that the probability the conclusion is true is very high. In evolutionary biology, however, conclusions can sometimes be made in the absence of data sufficient to even calculate a probability, or despite probabilities that are so vanishingly minuscule it is irrational to draw an inductive conclusion.
Example #1: Origin-of-Life Scenarios
Evolutionary biologist Eugene Koonin has argued that the probability of the emergence of RNA replication is so small, we cannot expect it to occur anywhere in the universe. Yet modern science has already committed to the foregone conclusion that blind and mindless natural processes created life. Recall that the validity of an inductive conclusion rests on its probability, given the data. The minuscule probability of RNA replication exposes this inference as devoid of rational justification, at least given the present data.
Example #2: Common Descent from a Simple Cell
Modern science is firmly committed to the belief that if we can just get life started, large-scale evolution is a foregone conclusion, leading to the huge diversity of life we see today. This would have required the origin of thousands of different protein families by blind and mindless processes. There are many creative stories in evolutionary biology that imagine how this could happen. Indeed, it is assumed by some to be inevitable. However, the actual, real-world data say otherwise. I am presently involved in a project that uses actual protein family data to estimate the probabilities of various protein families arising through blind natural processes. As an example, a protein family known as RecA is essential to all life. I used 9,170 RecA sequences to estimate an extreme upper limit for the probability of obtaining it without any intelligent input. The data shows it tolerates an average of 16 amino acids per site. If we assign equal weight to all amino acids (an extremely generous assumption), we obtain an extreme upper value for the probability of obtaining any sequence of RecA as 1 chance in 10 followed by 28 zeros. A more realistic estimate from the data is likely closer to 1 chance in 10 followed by 237 zeros. (My program and modules are available here, and the data for RecA are available here.)
This is just one protein family. The full range of biological life requires thousands of unique protein families.
The probability of getting mind-staggeringly lucky thousands of times over is so close to zero that there is no rational grounds for the inductive conclusion that large-scale evolution could happen under the blind and mindless “watchmaker” of natural processes.
In short, the trustworthiness of an inductive inference is founded on the probability or likelihood, given the data. The conclusion, therefore, that the information for the origin of life, and for the thousands of protein families required for life, arose through a blind and mindless “watchmaker” is so wildly improbable that it does not even begin to qualify as a rational inductive inference. So why does modern science put faith in a conclusion that reason says is utterly unjustifiable? The answer is the influence of scientism upon inferential science.
Scientism is the belief that science explains everything. This requires the a priori ruling-out of a mind behind the origin and diversity of life. Instead, everything must have a natural explanation, even if it cannot be experimentally reproduced and there is no rational justification for the inductive inference. Especially problematic is that, just as a woman cannot give birth to herself, it is logically impossible for the natural world to have a natural origin. Logic dictates that the origin of nature is not-natural, as I have shown in another post. The influence of scientism on science results in a “war on science.” Essentially, scientism is atheism dressed up in a lab coat. It is a philosophical position that has inordinately influenced otherwise good science. Consequently, science is forced to commit to conclusions before the experiments have even been done, and inferences that the data show have not the faintest likelihood of being trustworthy. Not surprisingly, most of the tension between faith and science arises from these types of scientism-driven inferences.
D. Ignoring Falsification of Key Predictions
Because so many conclusions in modern science depend upon inductive inferences to conclusions that cannot be proved experimentally, falsification has an important role. It tells us if we are on the wrong track. It is very bad science that focuses only on the positive support for a theory while ignoring experimental and observational evidence that falsifies it.
An essential prediction of the Darwinian theory of common descent, for example, is that functional genetic information increases through a process of mutations, insertions, and deletions. Experimental science, however, consistently falsifies this prediction. In reality, the number of harmful mutations is greater than the number of beneficial mutations, with the net result that the genomes of life are slowly degrading. We see this, for example, in bacteria, in the fruit fly, and in human beings. In this case, scientism’s philosophical commitment to common descent sets aside actual experimental results that contradict that belief because, under scientism, the foregone conclusion is required to be true, even if experimental science appears to falsify a key prediction. Scientism’s belief in Darwinian common descent by blind and mindless processes is, as some might say, “too big to fail.”
E. Inductive Conclusions Supported by “Lack of Data” Words
My PhD supervisor was carefully going over a paper I was about to submit for publication when he noticed a sentence in my conclusion that contained the word “suggests.” He asked if I had the data to support this. I replied that I did not — it was only an inference. “Then remove it,” he said. The word “suggests” was being used to substitute for a lack of sufficient data.
“Lack of data” words and phrases are rampant in Darwinian literature pertaining to scenarios for the origin of life and common descent by mindless processes. For example, in a short, two-page article, “The origin of the very first species and the start of Darwinian evolution,” I counted a total of 28 “lack of data” words including, “presumably,” “probably,” “possible,” “might have,” “at some time,” “possible scenario,” “could have,” “is conceivable,” “over time,” “researchers believe,” and so on.
The take-home point here is to begin to look for and recognize “lack of data” words in papers and articles where inferences are being made. When you see them, you are witnessing a transition into science fiction rather than data-justified inferences and good science.
On the positive side, there is much in inferential science that is very reliable, provided the inductive leaps are quite small, and the data to back up inductive moves are substantial enough to make them very probable or likely. The bigger the leap, however, and the more inductive moves required to arrive at the conclusion, the more skeptical we should be. Narratives about the history of the universe and of life can require a large number of inferences, some good and others not so well justified, not to mention an occasional huge inductive leap.
Not too surprisingly, many of the biggest tensions between faith in the Bible and faith in science are found within inductive conclusions where the largest and most numerous leaps are made, especially in discussions and papers on the origin and diversity of life through blind natural processes. We have good reason, therefore, to re-examine just how much faith we should be putting in certain inferences of science. We need to critically question current tensions between faith and some questionable inferences that are heavily influenced by scientism and those aforementioned “perverse incentives.” There are many solid, trustworthy inferences in modern science, but there are those that are not very trustworthy at all. The challenge is to discern the difference.
Photo: Hydrothermal vents, where some theories hypothesize that life originated, by NOAA [Public domain], via Wikimedia Commons.
Cross-posted at KirkDurston.com.