Optimal Design, Argument From
Optimal design, a term used commonly in engineering and science, describes a flawless, ideal and exemplary design. It is frequently invoked as central to arguments against intelligent design theory with its assertions that a designing intelligence is a better explanation for the complexity and specificity of biological systems, given the total absence of hard scientific evidence for chance based macro-evolution and the corresponding observable existence of irreducible complexity and mathematically rigorous probabilistic corollaries of complex specified information. Proponents of the optimal design criticism of intelligent design invoke optimal design as perfect design, claiming that any intelligence capable of designing biological systems would design them to be what modern, technologically advanced human beings regard as optimal and flawless1. The argument stipulates that nature is full of apparent and actual flaws, and so no designer - especially a perfect or extremely superior one - could be responsible for them. This argument is infelicitous because not only does it incorporate question-begging about the nature and intent of a putative designer, but it employs equivocation between the ideas of intelligent design and perfect design, and constructs a straw man view of intelligent design that obfuscates the real questions of whether intelligent design in nature is apparent or actual2.
To reveal the flaws in the optimal design based critique of intelligent design, one might begin by checking the underlying assumptions of its predicates. Firstly, it assumes that the designer of intelligent design theory is required to be perfect, inerrant and thus capable of and committed to producing perfect designs. However, intelligent design theory does not require that any designer of nature be perfect3. Furthermore, the argument from optimal design assumes that an intelligent designer will produce only perfect designs.An intelligent designer need not produce perfect designs - nor even functional designs. Optimal designs need not be perfect designs, and in fact it may be impossible for them to be so, depending on one's definition of perfect, and depending on whether such perfection is even materially attainable or desirable. Moreover, even a perfect intelligent designer (whatever or whoever that might be) need not produce perfect designs4. Many designs produced by human beings are both complex and specified, but implemented solely for aesthetic purposes. Historically, very intelligent people have often produced art and even industrial design for aesthetic purposes, or in some cases, even anti-aesthetic purposes. Moreover, an artist may be trying to communicate a message of some kind, but then different people will interpret the message different ways - not getting the intended message but some variation of it - or something completely different. This method of communicating a message could hardly be regarded as perfect - or optimal - from a mathematically rigorous or scientific perspective, such as that presented by Shannon's Information Theory. From the perspective of information theory, such art is hardly functionally optimal because, among other things, the 'signal' from the sender (artist) to the receiver is extremely noisy. However, an intelligent designer has been responsible for most of the art in history - even the abstract and postmodern art. In any case, we do not need to recourse to art for this illustration. Intelligent designers in science and engineering also often produce non-optimal designs.
Software engineering is a discipline that, although often referred to as a combination of art and science, has some of the most rigorous requirements for functional accuracy and optimal specification and implementation. This is largely because the computer systems on which software runs are not intelligent, and therefore cannot intelligently cater or adjust for human error, except on the basis of rules which have been built into them by (flawed) human intelligent designers. If mission critical software on the space shuttle fails due to a flaw, astronauts may be pu tat risk. If your 3D rendering software for your latest expensive blockbuster kids movie is not optimal(to your budget and schedule), then your movie will never get finished. However, most software developed contains flaws or errors of some kind - even mission critical software. Kludges (non-standard approaches to implementing functions in code), dead code (code that is there but does 'nothing') and spurious code (causing 'ghosts' in the machine) are just some of the flaws incorporated into high-end software programs by intelligent designers. In fact, software engineers often know where these flaws are - but choose to leave them in the source code (the actual human readable program) by design.
Why would a software engineer leave flaws in the program code? There are many possible reasons for this. Perhaps the offending portion of code, although imperfect, is known to be harmless, and the time and effort required to remove it, given budget constraints and the fact that the software has to be delivered to the client tomorrow, render the effort of removing the offending code and going through the lengthy process of re-compiling and testing insensible. Any intelligent observer who sees the code at a later date - perhaps five years after delivery - when the code needs to be re-furbished, will wonder how an intelligent software engineer could make such an error. A good example of this is dead code. Most modern software is developed using high-level, object-oriented programming languages. The term high-level refers to the fact that the
syntax and semantics of the language are such that humans can work with it more easily. High-level languages incorporate natural languages like English and standard constructs that allow modularity so that complex programming problems can be solved with a 'divide and conquer' approach. It would be impossible for the highly complex industrial and commercial software used today to be written solely in binary - or in low level languages like assembly language - which computers interpret readily, but which are very hard for people to work with and structure. Their mind-numbing repetitiveness and 'unnatural' syntax, which makes them very hard to structure and interpret, is prohibitive of human readability. Object-oriented refers to the constructs built into modern programming languages that facilitate modularity (divide and conquer), abstraction (hiding of details not specific to the immediate problem), hierarchy (new program objects inherit properties from existing program elements), and encapsulation (data is protected from code that should not touch it). Dead code is code that was written into the program in high level programming language, either erroneously, or originally, for some now redundant purpose (such as testing other code), and remains in the program source code doing nothing. Modern compilers, programs that convert source code into computer-interpretable code, usually strip out
such unused code during the process of compilation, analogous to the way in which DNA polymerase excises introns (non-coding sequences) out of a gene and splices together its exons (coding sequences) when transcribing to mRNA. However, the dead code can happily sit in the source code (as perhaps can the 'junk' DNA stay in the DNA) forever - doing nothing but causing confusion to later intelligent observers - engineers who must update the program for example.
Sometimes software engineers will leave code that is not dead code, but does not perform a function, in the source code intentionally. Such code is sometimes called a stub, and is incorporated for the purposes of extensibility - extending the capabilities of the code later on. The code is correct and required effort to produce, was deemed unnecessary to the current version of the software, but is desirable to add functions to the next version, and therefore is purposefully left in - by design. Of course, depending on a number of factors - whether requirements for the next version of the software change or other software makes it obsolete before the next version - even code for extensibility might never be used. There are many other examples of code, that might never actually be used, being purposefully incorporated into software programs. Take interface code for example. Software engineers might build software A to talk to or interoperate with software B written by other software engineers. The source code that facilitates this interoperability is called interface code. However, if the software B project is cancelled, then the interface to B from A will likely never be used, but might never be removed. This is just another example of designed flaws or non-optimal design by intelligent designers.
It is also noteworthy that in computer science and engineering, what is actually considered optimal is contingent on many factors relating to the requirements for the design (what people need or want it to do) and how powerfully the design must implement functions to cater for those requirements. For example, the speed of a software module that runs a screensaver is likely to be considered optimal at a much lower bit-rate than one that renders visual radar feedback on the space shuttle.
Ultimately, of course, software engineers are imperfect designers with limited capabilities operating in imperfect conditions (budget, fatigue and delivery schedules etc.), and so no rational person expects their designs to be perfect, flawless, or even necessarily optimal (although they may converge on some of these some of the time). However, even a putatively perfect or superhuman intelligent designer - one capable of designing the biochemical underpinnings of life - is arguably not required to produce perfect, flawless or optimal designs.
The objection, from optimal design, to the intelligent design hypothesis in nature is actually based on a theological argument - and an arguably erroneous one at that5. The underlying assumption is that an intelligence brilliant enough to have designed nature, must itself be in some way perfect, and its designs perfect. However, assertion is a theological one - reflecting the Judeo-Christian view of God. It is often the case that opponents of intelligent design assume that intelligent design theory is just theology (even specifically Biblical creationism) in another guise6. Because of this, the assumption is that the intelligence being postulated by intelligent design proponents is the morally and intellectually perfect Judeo-Christian God. This is a natural and intuitive conclusion given the initial assumption. However, intelligent design theorists are not attempting to assert the attributes, character, personality, attitude, features or configuration of an intelligent designer - only that somehow, somewhere, at some time, an actual intelligence of some kind was, and perhaps still is, involved in designing the material universe and the biological life in it7. The criticism from optimal design is in fact assumptive and circular, and a non-sequitur. If, as an
opponent of intelligent design, one asserts that the intelligent designer that an ID proponent proposes must necessarily be, for example, the perfect Judeo-Christian God, and that ID is therefore invalid because of the flaws in nature, which one further asserts could not have been designed by such a deity, then one has erred multiply in one's argument. Firstly, one has not only wrongly imputed to the ID proponent a belief in the Judeo-Christian God, but also imputed to ID an objective foreign to it - to assert the existence of said God. Furthermore, the argument has assumed that a perfect God would design perfectly. Any conclusions or inferences drawn on these base assumptions will be spurious.
As the previously indicated - there is another set of problems for objections based on optimal design. Firstly - even if one was referring to the Judeo Christian God, there is no reason why such a perfect designer must design perfectly and optimally. It would be completely the prerogative of such a being as to the degree of accuracy and flawlessness incorporated into their designs. Such a being could feasibly do whatever they wanted, in whatever way they wanted to do it, and what anyone else thought about it would be completely irrelevant. Although the reasons for such an intelligence to produce a perfect design are numerous and valid, the possible reasons for such a designer to produce what we regard as non-optimal design are perhaps infinite. Maybe such a being likes flaws - due to an aesthetic appreciation of the irregularities and a measure of chaos. Maybe said intelligent designer gets bored easily and once something works well enough, they are happy to leave well enough alone. This is the prerogative of the putative all-powerful supreme intelligence. Maybe everything is a work in progress. Maybe it is all a puzzle for other intelligences - sentient beings perhaps - to figure out. Maybe struggle and suffering due to flaws is somehow not the bad thing that human beings think it is, and no lesser sentient knows how. One could postulate forever. Of course, many of these options are not available to the believer in the Judeo-Christian God specifically, because of the well-defined moral and relational parameters of said God, and the nature of said God's character as asserted by Judeo-Christian doctrine. The Judeo-Christian God is constrained by theology (by himself according to theology) to care about creation, about human beings and suffering, and to be the implementer of an originally perfect, flawless design. This highlights another reason why the optimal design argument, which actually relies on theological assumptions, is not a problem for even intelligent design proponents who happen to be of Judeo-Christian orientation. Once in the realm of theology, Judeo-Christianity variously allows for imperfection in nature based on numerous doctrinal tenets, including the effects of the fall.
It is true that intelligent design allows far more scope for inference and investigation than methodological naturalism, because it is not circularly-inductively bound to the assertions of materialist ontology. In not being limited in its power and reach by naturalistic assumptions which stifle curiosity and limit scientific enquiry, intelligent design theory does allow for the proposition that the material universe was created by a deity - even a specific kind thereof, but does not provide any evidence for, nor seek to make any claims regarding, the character or attributes of any designing intelligence. In any case, this has never been the central pursuit of intelligent design. A more correct statement of the main objective of intelligent design might be that it seeks to unseat methodological naturalism, with its arguably scientifically untenable assertions that life is due to chance-based naturalistic Darwinian mechanisms only, as the cornerstone of modern science. Such an unhorsing of the battle-hardened, but flawed, warhorse that is the enlightenment interpretation of what science should be - methodological naturalism - is long overdue. That said - intelligent design theorists are not by any means seeking to unseat all of current scientific methodology, although this is a criticism frequently leveled by convinced macro-evolutionists and neo-Darwinians. Rigor, observation, repeatable experimentation, analysis based on evidence and conclusions based on hard science supported with rigorous mathematical analysis are essential to the intelligent design view of science. It is rigid, closed materialism and the assertions of methodological naturalism - that complex biological systems display only apparent design as the product of chance events over time -that is the object of the incredulity of intelligent design. Optimal design remains a possibility in concert with the possibility of a designer perfect enough to achieve it.
1.William Dembski. The Design Revolution: Answering the
Toughest Questions About Intelligent Design, Downers Grove,
Illinois: Inter Varsity Press, 2004, 57.
2. Ibid., 58.
3. Ibid., 58.
4. Ibid., 59.
5. Ibid., 62.
6. Ibid., 45.
7. Ibid., 42.
Dembski, W. Taken from The Design Revolution by William A. Dembski.
© 2004 by William A. Dembski. Used by permission of InterVarsity
Press, P.O. Box 1400, Downers Grove, IL 60515-1426. www.ivpress.com
Web Resources On Optimal Design, Argument From
Editor(s): Long, B.