Personal Explanation in Natural Theology and Intelligent Design

I have argued that ID attempts to explain the presence of teleological features of nature in terms of the operation of a mind. Precisely what sort of features is meant will become clear in chapters five and six. For now, there are even more fundamental questions to discuss: how do explanations by reference to the intentional activity of an agent work? Or do they have explanatory value at all? There is a vast history to this discussion.²⁶¹ This chapter will introduce the discussion on the validity of design arguments which will be continued in the following chapters. I will also note some commonalities and differences between ID and theistic natural theology.

Personal Explanation

In the contemporary discussion, the concept of personal explanation has received particular attention in relation to theistic natural theology. Because of the success of natural science, mechanistic explanations involving reference to law-like regularities are often considered to be the paradigm case of good explanations.²⁶² But there is another type of explanation that all humans are familiar with: personal explanation by reference to the purposes and capabilities of agents. The possibility that such explanations can have explanatory power is assumed in normal human life whenever we explain the activities of some human person by reference to their intentions. For example, we might say that our friend went to the shop because she wanted to buy milk, or we might say that an automobile has the order it has because it has

258 Ruse 2003. Thus also Dennett (2006b, 38) argues that design in nature is real and ingenious, but is explained by a process which acts without intelligence.

259 Dawkins 1991, 1. See chapters 2.2 and 4.1 of the present study for further discussion on ID´s use of Dawkins.

260 Ruse 2003.

261 Knuuttila & Sihvola 2014 for in-depth discussion.

262 See Salmon 1990 for discussion.

been intentionally designed by humans. Similarly, theistic natural theologians argue that nature has certain features because its order has been intentionally designed by God.²⁶³

The ID movement does not always specify very exactly what is meant by design as an explanation, but the basic idea of personal explanation is clearly visible.²⁶⁴ Dembski makes the distinction between “design” which refers to just the pattern or structure which is observed and “intelligent design”, the activity of an intelligent agent which can create such a pattern. Dembski argues that the process of intelligent design typically proceeds in three steps. First, the designer has a purpose which he wants to execute. Second, the designer forms a plan to fullfill his purpose. Third, the designer arranges things according to his purpose. ²⁶⁵ Other ID theorists argue that to create certain types of structures, the capability of an intelligent designer to arrange parts with foresight is required.²⁶⁶ Principally, the designer is capable of “directed contingency”, in other words “choice”.²⁶⁷ The designer thus has a will, the ability to plan, and the ability to influence the material world. The definition is based on human experience of design, just as the natural theologian’s defence of divine personal explanation is based on the efficacy of human personal explanation. The argument of the ID movement is that designers can create rational patterns in nature. Human designers can do this, and similar (but greater) capabilities are seen as the best explanation for nature’s order.²⁶⁸

Many objections to the use of personal explanations in natural theology and Intelligent Design thought have been made. At the most basic level, the coherence of the concept of God could be questioned. If the concept of God was somehow contradictory or incomprehensible, then explanations based on the activity of this divine person would not be explanations at all. Thus atheist Kai Nielsen, for example, has argued that the concept of God is so incomprehensible that it would not be reasonable to believe in God even if the stars were arranged before our eyes into the text “GOD EXISTS”, because this would have no more meaning that if the text said “PROCRASTINATION DRINKS MELANCHOLY.”²⁶⁹ ID is not committed to seeing the Creator as divine. However, if the concept of an unembodied intelligence is incoherent, this would be a problem for any form of Intelligent Design which implies that the designer is unembodied. In chapter 3.1, I argued that this is indeed how the ID theorists in practice understand the designer, so these arguments are also important for them.

In defending the coherence of theistic explanations, natural theology requires the help of philosophical theology. The discussion started by Swinburne’s “The Coherence of Theism”

263 E.g. Swinburne 2004.

264 Menuge 2004b.

265 Dembski 2002, xvi.

266 Behe 2006a, 193.

267 Dembski 2002, see also Dembski & Witt 2010. Dembski expands on his definition by writing that “[t]he principle characteristic of intelligent agency is directed contingency, or what we call choice.” (2002, 62)

268 I will be analyzing the logical structure of the argument in more detail in chapter 4.

269 Nielsen 2004, 279.

(1977) has convinced many of the coherence of theism and theistic explanation.²⁷⁰ A popular way of arguing for the coherence of theistic personal explanation is by reference to human personal explanations. According to Charles Taliaferro, if we regard explanations in terms of human intentions (such as “I wish to see my beloved”) as explanatory of human behaviour and logically coherent, then we will also have to concede that theistic personal explanation is also coherent and explanatory.²⁷¹ However, others continue to argue that the coherence of theism has not been demonstrated. The divine person is also often thought to be difficult to describe using the same concepts as human persons.²⁷²

Another simple way of rejecting theistic arguments would be to also reject the coherence and reliability of intentional explanations even in the case of humans. Eliminative naturalists, who believe that mind and intention are not real explanations even in the human context, will naturally also reject biological and cosmic design arguments. Within eliminative naturalism, references to intentionality in the case of humans are seen as “folk psychology”

which should be replaced with correct scientific neurophysiological explanations for human behaviour.²⁷³ However, eliminative naturalism is a highly problematic view, which many theistic thinkers even regard as the reductio ad absurdum of naturalism, if it really is the most consistent naturalistic view of the world.²⁷⁴ In the discussion on ID, naturalists have typically not embraced eliminativism, but have accepted the validity of design-based explanations in the human context. Thus references to design and purpose are typically accepted to have explanatory power in the human context, but applying the inferences to the order of nature is criticized. This is a much more promising strategy than eliminative naturalism. However, it also allows that if design-based explanations used in biology are sufficiently similar to the explanations used in the case of human agents, then they could succeed.

Evaluating Design Arguments

The precise ways of criticizing ID’s design inference will become clear in the coming chapters, but a preliminary list of criteria – testability, the amount of details, relation to background knowledge and past explanatory success – will be helpful at this point to elucidate differences between personal explanations and mechanistic explanations. In the discussion surrounding ID, design-based explanations are typically evaluated using criteria developed in the philosophy of science to evaluate different scientific hypotheses. Science is held as the gold standard to which other explanations are compared.²⁷⁵ This strategy is not wholly misguided, since the ID theorists themselves present their design arguments as part of the

270 Swinburne 1993.

271 Taliaferro 2012, 8-16; Taliaferro 1994.

272 Dawes 2009, 48-51.

273 E.g. Churchland 2007.

274 E.g. Feser 2008, Goetz & Taliaferro 2008, Cunningham 2010, Menuge 2004b.

275 My list is based on criteria discussed by Ratzsch (2001, parts III and IV) and Dawes (2009, chapter 7).

natural sciences. Furthermore, there is surely much common ground between the evaluation of design as an explanation includes reference to criteria that are used in the natural sciences. Nevertheless, ID as a personal explanation also differs substantially from explanations that are typically used in the natural sciences.

The first commonly referred criterion of good explanations is the criterion of testability.

In the natural sciences, the possibility to test different ideas against empirical results is highly valued, since otherwise we would be left with no way to find out if our theory is false.

Theories can be tested by reference to their ability to explain current scientific knowledge, but also by their ability to predict future results. In some cases, predictions can be fairly easily inferred from the postulates of the theory. For example, knowing the laws of thermodynamics, we can predict that the amount of usable energy in a closed system will decrease over time. But often auxiliary hypotheses have to be added to the theory before we can derive predictions from it.²⁷⁶

Following the ideal of testability, it can be argued that personal explanations are also testable. For example, let's hypothesize that a person is going to the shop to buy milk. We can look at what the person does, and notice if the result actually occurs.²⁷⁷ Intentional explanations can also help be tested by comparison to other explanations, like the null hypothesis (random chance). Supposing that humans act for reasons provides a more plausible overall picture of human activity than the hypothesis that they act for no reason at all. Supposing that there exists a being with the appropriate intention and causal power to create some result, it can be argued that this hypothesis predicts the data better than the null hypothesis.²⁷⁸

However, typically, our theories about agents are not formed in this manner, by first hypothesizing about specific motives and then observing human behaviour. As Ratzsch notes, “design theories are, ultimately, theories involving agency. And, with respect to agents, theories and explanations often must trail data. We are often in positions of utter inability to predict specific human actions, but such actions once observed may be readily and quite legitimately explainable.”²⁷⁹ This is the case even in the case of humans. For example, we may be unable to predict why a person is going to the shop. Is it to buy milk, or to buy sugar? Or perhaps to flirt with the pretty clerk? We are often unable to tell beforehand. Often we even need information from the person in question before we can make the choice. Reasons for behaviour are often easier to identify after the fact, rather than beforehand.

In the case of ID, the motive of the designer is not thought to be part of the explanation.

Rather, the capabilities of the designer to produce goal-directed order are the important explaining feature. The ID movement has also attempted to derive other predictions from the design hypothesis, as we will see in chapters 5 and 6. I will argue that this is moving between a minimalistic and a more robust design hypothesis. In principle it is indeed possible to add

276 Ratzsch 2001, chapter 7.

277 Dawes 2009, appendix.

278 Dawes 2009, 125.

279 Ratzsch 2001, 117.

auxiliary hypotheses to a minimalistic design argument to make the hypothesis more robust and generate more predictions. If the initial design argument is successful in identifying that there indeed exists an intelligent designer, it does not seem unreasonable to try to make additional hypotheses about any patterns that can be discerned in the designer’s activity, just as extended observation of a human agent can lead us to create a more detailed theory about how that agent tends to act.²⁸⁰ So, as it keeps trailing the data, a design hypothesis can perhaps become more robust and thus also more testable, even if it is initially quite vague.²⁸¹

In chapter 4.4, I will be returning to the question of whether design hypotheses need to be predictive in order to be testable. In the influential hypothetico-deductive model of scientific explanation developed by Carl Hempel, explanation and prediction were synonymous. It was thought that each proper scientific explanation was also a prediction, because based on the relevant natural laws and conditions, one could predict the result. But in contemporary philosophy of science, the hypothetico-deductive model's strong link between predictive and explanatory capacity has been denied. Scientific explanation is more concerned with understanding the causal structure of the universe and explanatory unification, rather than just prediction. For example, quantum physics can explain even highly improbable quantum fluctuations, but may not ever be able to predict them.²⁸²

The amount of details provided about the designer is also relevant for analysing how good design is as an explanation. The goodness of God is a central explanatory factor in many theistic arguments. If we are able to comprehend God's goodness at least in some way, then it becomes possible to argue that a good God would have the motive to create a certain kind world. Thus Swinburne argues that God explains the order of the universe, because it is plausible to believe that his goodness would lead him to create bodily creatures capable of interacting with each other.²⁸³ Robin Collins argues that if we assume that the existence of complex, intelligent life is a good thing, then it becomes plausible to assume that a good God would create a cosmos which makes this existence possible.²⁸⁴ However, for many skeptics, the problem of evil makes ascribing such comprehensible goodness to God difficult.²⁸⁵ By contrast, the ID movement’s designer is left unidentified and his motives are thus

280 Dawes 2009, 266.

281 William Lane Craig (2007) has argued that although one cannot falsify the existence of a designer, particular forms of the design hypothesis can nevertheless be falsified and tested in this manner. It seems to me that this is correct. For example, one could falsify the hypothesis that the designer has created biological machines that are not evolvable through a gradual Darwinian process. Indeed, as we will see in chapter 6.3, many argue that natural science has indeed falsified this form of the design argument.

282 See Salmon 1990, 117-122 and Woodward J. 2011 for the history of this discussion, as well as Woodward J. 2003 for discussion of the concept of causality. Cleland (2011, 568-569) argues that this can be seen even in the historical natural sciences. For example, the hypothesis that an asteroid impacted the Earth 65 million years ago arguably was not able to predict the precise geological discoveries that we have made, but is nevertheless regarded as the best explanation.

283 Swinburne 2004a, 346-349; 2004c.

284 Collins 2012, 254-256.

285 Narveson 2003. Similarly, Dawes (2009, 44-46) argues that a hypothesis about the motivations of the deity is extremely important for the explanatory value of theism.

mysterious. As I will argue in the coming chapters, the ID movement argues that certain types of patterns are better explained by an intelligent cause, even if the motives of this intelligent designer are unknown.²⁸⁶

Evaluations based on the amount of details provided by an explanation are also related to the concept of scientific tractability. Scientific explanations are highly detailed, with references to laws, mechanisms and the minutest details of the systems being investigated.

Science also does not provide explanations for everything – for example, it is difficult to specify a mechanism explaining why gravity works the way it does. However, in general, an attempt is made for investigating natural phenomena in detail. By contrast, explanations by reference to intentionality do not include this level of mechanical detail. Theistic intentional explanations typically do not involve any specification about the mechanism by which God creates the laws of nature, for example. Indeed no such mechanism needs to be given, since according to the hypothesis God can bring about any result he chooses without any need for intermediate second causes. In the case of the intelligent design hypothesis, an intermediate cause does exist if the designer of life was a space-alien, as has been suggested. However, nothing like this has been specified by the Intelligent Design movement, and no evidence of laboratories or similar has been found on Earth.

Both theistic natural theology and Intelligent Design require that intentional activity as a cause itself possesses some explanatory power, even without specification of any particular mechanistic process the designer worked through. In intentional explanations as used in the human context, this does seem to be the case. In all explanations, there comes a point where we reach the level of basic causal powers, and are unable to specify further intermediate mechanisms.²⁸⁷ Furthermore, while intentional explanations seem to work on a different level than mechanistic explanations, this does not mean that they have no explanatory power.²⁸⁸ Like the natural theologians, ID theorist Meyer also references the example of human design as one basis or this claim. In the case of humans, we cannot yet specify how our consciousness and will influence the molecules of our bodies, but we nevertheless believe that our consciousness has an effect in the world, and that references to human design can be

286 This is a crucial difference between ID's design argument and the intentional explanations described by Dawes (2009, appendix). Within ID, the capabilities of the designer are thought to have explanatory power even apart from knowing the precise motives of the designer.

287 There is reason to think that the demand for intermediate mechanisms should not be absolute. Dawes (2009, 51-53) argues that the demand of an intermediate mechanism leads to an infinite regress. Suppose that we cannot say that A causes B without specifying an additional intermediate mechanism C. How can we then say that A causes C to cause B without specifying yet further causal mechanisms for how A causes C, leading to an infinite regress? Dawes thinks that divine action is a mysterious concept far removed from our everyday understanding, but the requirement for a mechanism is no reason to reject theistic explanations. At most it can be a reason for preferring naturalistic explanations when such are available.

288 Dawes 2009, appendix. These features lead Collins (2006) to suggest that Intelligent Design should be viewed as a ”metascientific” idea: it has empirical content, yet operates on a different level than normal scientific theories.

I will return to the question whether Intelligent Design is science in chapters 3.4 and 3.5.

explanatory. In addition, we can typically detect that something is designed by humans without being able to specify how these humans do so. ²⁸⁹

This does show that design can be explanatory even without precise knowledge of mechanisms. However, in the case of human designers, we do typically have at least some idea of how the designed objects were produced (or how they could have been produced) in practice. The possibility to investigate such details further is a good thing for the hypothesis, though it is not unconceivable that a hypothesis could not have explanatory power even if further details about the cause cannot yet be investigated. ²⁹⁰ If personal explanation is indeed explanatory, it is not necessarily a fatal flaw for it that personal explanations are different from mechanistic explanations. Even a vague hypothesis could in principle be the most plausible one, and could provide us with valuable knowledge. Following Aristotle, it could be argued that there may be great value in even a glimpse off “celestial things”: “half glimpse of persons we love is more delightful than an accurate view of other things.”²⁹¹ But such

This does show that design can be explanatory even without precise knowledge of mechanisms. However, in the case of human designers, we do typically have at least some idea of how the designed objects were produced (or how they could have been produced) in practice. The possibility to investigate such details further is a good thing for the hypothesis, though it is not unconceivable that a hypothesis could not have explanatory power even if further details about the cause cannot yet be investigated. ²⁹⁰ If personal explanation is indeed explanatory, it is not necessarily a fatal flaw for it that personal explanations are different from mechanistic explanations. Even a vague hypothesis could in principle be the most plausible one, and could provide us with valuable knowledge. Following Aristotle, it could be argued that there may be great value in even a glimpse off “celestial things”: “half glimpse of persons we love is more delightful than an accurate view of other things.”²⁹¹ But such