Recently I engaged in discussion with Dr. Stephen Meyer, one of the world’s foremost advocates of the theory of Intelligent Design (ID). ID, at least as Meyer presents it, argues that certain features of the natural world are indicative of a creative intelligence: he focuses on the origin of the universe; the so-called “fine-tuning” of the universe for life; and the origin of life. These three phenomena, he argues, strongly suggest the existence of a designing intelligence with the qualities traditionally ascribed to God. You can find our discussion here.
In general, I was happy with our conversation. I feel like I presented some significant challenges to Dr. Meyer’s view. I was frustrated by a couple of technical glitches, and somewhat miffed by the amount of airtime Dr. Meyer took up (he had by far more time to speak than I did), but in general I thought it was a genial and valuable exchange. If online engagement is any guide then discussion was a success: it has almost ten times as many views as any other video on the host’s channel, and has provoked a spirited discussion in the comments, of which there are at the time of writing 150. This discussion has given rise to some common questions and misconceptions, however, and here in this (quite lengthy) post I want to address them.
Meyer’s argument does not involve a proof that natural processes cannot give rise to complexity
The first misconception regards Dr. Meyer’s basic argument. Because Intelligent Design is not a single argument, but a movement or intellectual tendency which admits of many variations, some people have become confused about what Meyer is arguing. Some ID advocates argue that phenomena such as life exhibit certain properties which can, in principle, only be explained by the actions of an intelligent agent. William Dembski, for instance, tries to show through mathematical reasoning that an “undirected process” cannot give rise to “complex specified information” (a property which, he argues, we only find in things which are the result of design by intelligent agents). If Dembski’s argument is correct, and we were able to show that any phenomenon displayed “complex specified information”, then we would be logically required to conclude that it was not the result of an undirected process. He would have proved the inability of undirected natural processes to give rise to things with that property, and that only intelligent agents can do so, so we would have no other option but to conclude that intelligent agents were the cause.
There are all sorts of problems with Dembskian ID arguments. His notion of “complex specified information” has been totally rejected by the mainstream academic community, and there are numerous detailed examinations of the mathematics which underly his theory showing that he makes profound mathematical and philosophical mistakes. In my view, the Dembskian project of proving that a process such as evolution cannot, in principle, give rise to the life that we see has been utterly dismantled. Some reasons to believe this are that 1) hardly any mainstream scientists make use of his ideas; and 2) ID proponents like Meyer do not rely on his arguments. But my judgment of a Dembskian attempt to prove that evolution can’t give rise to biological complexity is not so important here: what’s important is that this is not the sort of case Dr. Meyer makes.
Dr. Meyer, by contrast to someone like Dembski, does not argue that we can prove that an “undirected process” cannot give rise to complex things like life. Perhaps he thinks such an in-principle argument might be made (he gestures in that direction in our discussion), but he doesn’t offer one in his books and public presentations on Intelligent Design. Instead, he offers an “inference to the best explanation.” Rather than saying “We can prove that natural processes could never give rise to life” (for instance), he says “Intelligent agency is a better explanation for the origin of life than a natural process.”
This is a critical distinction that some commentators fail to make. It’s important for a lot of reasons, but mostly (I think) because it changes the status of ID proponents’ critiques of current science. If you can prove that natural processes cannot give rise to phenomena with property X, then you know every time you encounter a phenomenon with property X that it was not the result of a natural process. In such an argument, the critique of current scientific explanations is definitive, because they show (if the argument is successful) that the phenomenon under discussion simply could never have arisen by a natural process. But under a Meyerian inference to the best explanation, you have never proved that natural processes could not, in principle, give rise to phenomena with X. There is always the possibility that a currently unknown natural process could be discovered which explains how phenomena with X come about. So the question becomes not just “Does this phenomenon display X?” but “Are phenomena with X better explained by appeals to an as-yet unknown natural process, or by appeals to an intelligent agent? Or should we just accept our ignorance and work on better explanations?” This is a very different set of questions, and they need to be approached differently.
Meyer’s argument for design
So why does Meyer think that the origin of life, for instance, is best explained by an appeal to a designer (I focus on his argument for design in biology, rather than in cosmology or physics, both because it is main area of interest and because it is much easier for laypeople to understand. The cosmological arguments get complex fast)? Because 1) he does not believe current scientific theories can explain the origin of life and 2) life displays properties he claims are only seen in other things we know to be the products of intelligent agents. So, by combining these two observations, he thinks the most reasonable hypothesis we can propose is that an intelligent designer is responsible for life. Expressed as a syllogism, Meyer’s argument is as follows:
Current scientific theories are unable to explain, and are unlikely to explain, certain observed phenomena, such as the origin of life.
All these phenomena exhibit characteristics which are also and only found in phenomena we know to be the result of intelligent agency. Therefore,
These phenomena were most likely also the result of intelligent agency.
Note that 1. does not involve a proof that no scientific will ever be able to explain the phenomena in question, so Meyer keeps open the possibility that some future theory might do so. He just note that none has yet, and argues that the most current science renders it unlikely that one will.
Note that 2. appeals to a similarity between things we know to have been the product of intelligent agency (he gives examples like computer code and radio communications) and life, and suggests that because both share this particular feature, it is reasonable to think they had a similar cause. The feature Meyer picks up on is a variant on Dembski’s “complex specified information” which Meyer calls “complex functional information”, but it doesn’t really matter what it is because, as I hope to show, it’s an illegitimate inference in any case unless you can show that the feature is only in principle a feature of designed things – something Meyer does not do and does not claim to do.
Having clarified the nature of Dr. Meyer’s argument, I have a few things to say about it to expand upon some of the points of contention in our debate. My main problem with the argument Meyer presents is that it seems to me an illegitimate inference based on the misapplication of epistemic principles.
The problem of background knowledge
First, as I explained in my opening statement, the abductive inference Meyer draws is extremely unusual, in that it suggests that the cause of an observed effect is a designer about which we have no background information. When we normally make an abductive inference to a designer – reasoning that a human being was responsible for breaking into our car, for instance – the inference is secure because we have a lot of background knowledge about the proposed agent. We know that human beings exist; we know that they have the causal means necessary to break into cars; and we know that they have the motive to do so. With all that background information already in place, we are able to reason back from an observed effect (our car being broken into) to a plausible agent (a human being).
The God Hypothesis (to use Meyer’s preferred phrase) is unusual in that it attempts to infer back to the actions of an agent for which we have no background knowledge at all (at least in our everyday sense of “background knowledge”). We do not know that God exists prior to running the inference: indeed, the phenomena to be explained (the explanandum) is supposed to be part of our evidence that God even exists. We have not established using independent evidence that God has the causal means to bring about the phenomena. And we have no independent evidence regarding God’s motives either. So unless we come to the problem with pre-existing reasons to believe God exists and has certain powers and motives, we seem to be in the unenviable position of trying to show that our explanans exists at the same time as trying to use the explanans to explain the explanandum. It seems to me that the evidence we have is far insufficient to do this. That’s the problem of background knowledge.
The Uniformitarian Principle is misused
But this isn’t the only epistemic problem with Meyer’s argument. In his book and other writings Meyer frequently refers to the uniformitarian principle, an epistemic guideline employed by geologist Charles Lyell. Specifically, Meyer refers to the subtitle of Lyell’s book Principles of Geology: Being an Attempt to Explain the Former Changes of the Earth's Surface, by Reference to Causes Now in Operation. Meyer calls attention to the phrase “causes now in operation” and argues that, like Lyell in his geological work, when we are faced with a phenomenon of unknown origin (like life), we should appeal to causes which we know to be in operation now to explain the phenomenon rather than posit entirely new causes we don’t know anything about.
Putting Lyell in historical context helps here. Lyell was engaged in an intellectual dispute with a group of scientists known as “catastrophists,” who argued that certain geological features of the world (which at that time had not been satisfactorily explained) were the result of sudden “catastrophes.” Under this view, the earth in the ancient past was periodically riven with violent earthquakes which created mountain ranges, or was deluged with great floods. The earth was mainly stable, but underwent these periodic upheavals which radically changed its features. “Uniformitarians” like Lyell, on the other hand, thought that the present features of the earth are the result of the same physical processes we see operating today, just built up over huge amounts of time. He rejected the idea that there were periods in the past when catastrophic forces operated which we no longer observe: instead, if we want to know how a mountain range or a canyon came to exist, we should look for “causes now in operation” and extrapolate them backwards.
There is a lot which could be said about this historic intellectual disagreement. Suffice to say that today, geologists understand that the features of the earth are the result of a combination of very slow, gradual processes, and some relatively fast “catastrophic” ones. But as an epistemic matter, the “uniformitarians” decisively won the debate: the principle that the laws of physics are uniform across time, and that we shouldn’t try to explain things which happened in the past by appeal to “exotic” causes which were in operation only then – is the standard position of scientists today.
Meyer claims to be working within this uniformitarian tradition, and explicitly invokes Lyell’s words to support his inference to God. But is he really referencing “causes now in operation” to explain the origin of the life and the universe, or is he proposing an exotic, unknown cause as did the catastrophists? The question answers itself, surely: Intelligent Design is a distinctly catastrophist endeavor. Rather than believing that the causes now in operation are sufficient to explain the origin of life, Meyer and his colleagues hypothesize the existence of a cause for which there is no independent evidence; which has unlimited causal powers; and which has curious properties (such as a disembodied mind and non-physical causal agency) unlike anything else we observe.
It is of course evolutionists who are carrying on the tradition of Lyell and the uniformitarians by suggesting that life arose gradually through some combination of known physical processes. Indeed, Darwin was influenced by Lyell’s thinking, and his theory of evolution is the paradigmatic application of the “causes now in operation” principle to help explain an hitherto unexplained phenomenon. Lyell would have been proud: he once wrote, of his own geological work, “Many appearances, which had for a long time been regarded as indicating mysterious and extraordinary agency, were finally recognized as the necessary result of the laws now governing the material world”. This is the uniformitarian hope, and a successful application of these principles would explain the origin of life, as well as the origin and fine-tuning of the universe, by reference to “laws now governing the material world”, and not to the “mysterious and extraordinary agency” of a God.
Meyer, then, is misapplying uniformitarian epistemic principles, draping himself in a mantle which is rightfully ours. If you want to be a uniformitarian, be an evolutionist; if you want to be a catastrophist, be an ID proponent.
“Intelligence” is not a “cause”, and humans are the only intelligent agents we have observed
One obvious response ID proponents might make to this critique is to say that they are appealing to a “cause now in operation”: intelligence! Indeed, this is exactly what Meyer argues, saying that because we see intelligences creating things which share specific characteristics with, for instance, life, then it is reasonable to ascribe life to that same “cause.” There are tons of problems with this approach. First, “intelligence” is not a cause. “Intelligence” is an abstract concept, a descriptive term we use to describe certain behaviors or traits of animals. “Intelligence” doesn’t do anything on its own: it is specific intelligent agents which do things, and we give the label “intelligent” to them or their behaviors based on the sorts of things they do. So you cannot say that “intelligence” is the cause of something, and then reason that “intelligence” was also the cause of something different-but-similar, because strictly speaking “intelligence” doesn’t cause anything.
OK, says the ID proponent, then I will propose that “intelligent agents” are the “cause now in operation” and that we should reason that “intelligent agents” are the cause of biological information. I don’t think this rescues the inference, because right now we have observed exactly one type of intelligent agent producing anything, and that is human beings. That means that Meyer cannot really say that intelligent agents in general are the cause of phenomena with “complex functional information”: he can only say that human beings have been known to produce such phenomena. And even if only human beings have been seen to do so, that hardly legitimates the more general claim that other, non-human intelligent agents will.
If Meyer were able to point to at least one other intelligent agent which has produced “complex functional information”, then he might be able to establish that intelligent agents in general tend to produce CFI, but with only one example he simply cannot. God, for Its part, is so unlike human beings in numerous significant ways that it seems very tenuous to say that because humans have been observed to produce phenomena with CFI, then it is reasonable to believe that God would too.
For these two reasons, then, I do not think it is accurate to say that to say that in appealing either to “intelligence” simpliciter, or to “intelligent agents”, ID proponents are mobilizing our knowledge of “causes now in operation” to explain an unexplained phenomenon. “Intelligence” simpliciter is not a cause, and there is no class of “intelligent agents” to which they can appeal – there is only human beings. They are actually appealing, I suggest, to the sort of “mysterious and extraordinary agency” Lyell decried.
“Retrospective Causal Analysis” is misapplied
In addition to the uniformitarian principle, Meyer frequently appeals to an epistemic practice he usually called “reconstructive causal analysis” to justify the design inference he wishes to draw. This is a reference (at least in part) to the work of philosopher Michael Scriven, who in fact called this mode of inference “retrospective causal analysis” (it seems somewhere in between Meyer’s first and third books the name got changed, and he hasn’t gone back to the original source and noticed his error).
Meyer suggests that this form of analysis legitimates the inference he wishes to draw, because it enables us to “look back” from causes now in operation to “reconstruct” what might have happened in the past. We know that, in the present, intelligent agents are the sole known source of information (Meyer argues); there is information necessary for life and for the universe to come about; therefore, we can reasonably conclude (absent any material explanation) that intelligence was the cause of this information too.
Scriven’s actual point is relatively complex in comparison with Meyer’s usage: Scriven argues that, given a certain data set, we may not be able to predict a certain phenomenon will come about; but, if we do observe that the phenomenon has come about, we can legitimately infer that elements of the data set were part of the cause (we can “reconstruct” a cause even if knowledge of the cause prior to the phenomenon coming about would not have enabled to predict that it would do so).
He gives the example of a fisherman who develops skin cancer: fishermen work a lot in the sun and suffer many skin abrasions; we know that spending a lot of time in the sun and suffering repeated skin abrasions are predictors of developing skin cancer; however, most people in that profession nonetheless don’t develop skin cancer. Therefore, just given the information that someone is a fisherman, we cannot predict that they will develop skin cancer, and indeed the safest prediction is that they will not. However, if a given fisherman does develop skin cancer, we can confidently conclude that the time in the sun and the skin abrasions are a cause. We assign the cause retrospectively, even though they did not enable us to predict the future. That is Scriven’s point: he is talking about the relationship between prediction and causal assignation (Scriven, 1959, p. 480 for the example).
In the very same paper, Scriven explicitly cautions against the unmotivated construction of explanatory hypotheses without independent evidence, however. He writes:
“Careless use of such arguments does produce ad hoc explanations… We have to show, as in the cancer case, that (i) this cause was in fact present, (ii) independent evidence supports the claim that it can produce this effect, and (iii) no other such causes were present. That this can be done is the mark, and a well-earned mark, of success” (p. 481)
Scriven here stresses that in order to engage in “retrospective causal analysis” of the sort he is describing, we have to have independent evidence that the cause we are appealing to was in fact present, can actually do what we want it to explain, and is not out-competed by other explanatory hypotheses. Meyer argues that we can meet the third criterion for his ID proposal (he argues that no other causes for fine tuning and the origin of life/the universe are present), but he ignores the first two: independent evidence that the cause to which he appeals (God) was in fact present and can produce the desired effects.
What Scriven is saying, essentially, is that in order to retrospectively assign some cause to a phenomenon, and thus (at least partially) explain that phenomenon, it needs already to be part of our known causal repertoire. Tailoring a hypothetical unknown cause to fit an observation is ad hoc: it is unmotivated by any broader epistemic considerations. Yet this is exactly what Meyer is doing.
That’s enough for now. I think I’ve shown that Dr. Meyer is misapplying the epistemic principles to which he is appealing, and therefore he is not able to draw a secure or convincing inference to intelligent agency on the part of God. For this reason, he has also failed to make a convincing case for design.