1. Computer scientist and philosopher Amnon H. Eden (Eden, 2007) seeks to bring clarity to the science-vs.-math-vs.-engineering controversy by taking up a dis- tinction due to Peter Wegner (1976) among three different “Kuhnian paradigms” (see Ch. 4, §4.9.2): a view of CS as (1) a “rationalist” or “mathematical” dis- cipline, (2) a “technocratic” or “technological” discipline, and (3) a “scientific” discipline. (Tedre and Sutinen 2008 also discusses these three paradigms.) Eden then argues in favor of the scientific paradigm.
But must there be a single paradigm? Are there any disciplines with multiple paradigms? Does the existence of multiple paradigms mean that there is no unitary discipline of CS? Or can all the paradigms co-exist?
2. Journalist Steve Lohr (2008) quotes a high-school math and CS teacher as saying, “I do feel that computer science really helps students understand mathematics . . . And I would use computers more in math, if I had access to a computer lab.”
Is CS best seen as the use of a physical tool, or as the study of (as well as the use of) a method of thinking (“computational thinking”)?
3. The philosopher Gottfried Wilhelm Leibniz (1646–1716) thought that alingua characteristica universalis(or universal formal language) and acalculus ratio- cinator(or formal logic) would offer “mankind . . . a new instrument which will enhance the capabilities of the mind to a far greater extent than optical instru- ments strengthen the eyes” (Leibniz, 1677, p. 23). From this statement, com- puter scientist Moshe Vardi (2011a) derives a “definition of computing, as an ‘instrument for the human mind’.” This is similar to Daniel C. Dennett’s sugges- tion that the computer is a “prosthesis” for the mind (see, for example, (Dennett, 1982)).
Is that a reasonable definition of CS?
4. In§3.12, I said that it makes no—or very little—sense to have a program without a computer to run it on. That a computer is useful, but not necessary, is demon- strated by the “Computer Science Unplugged” project (http://csunplugged.org/). And some of the earliest AI programs (for playing chess) were executed by hand (Shannon 1950; Turing 1953; https://chessprogramming.wikispaces.com/ Turochamp).
So,didthese programs “have a computer to run on”? Were the humans, who hand-executed them, the “computers” that these programs “ran on”? When you debug a computer program, do you do the debugging by hand?32
5. Forsythe observed that,
in the long run the solution of problems in fieldXon a computer should be- long to fieldX, and CS should concentrate on finding and explaining the prin- ciples [“the methodology”] of problem solving [with computers]. (Forsythe, 1967b, p. 454)
Should contributions made by AI researchers to philosophy or psychology be considered to be the results of AI? Or are they philosophical or psychological results that were onlyproducedorfacilitatedby computational techniques? 6. Maybe when Knuth says that CS is the “study” of algorithms, by ‘study’ he
meansbothscienceandengineering. In what sense does the study of electricity belong both to engineering and to science? Certainly, the science of physics stud- ies electricity as a physical phenomenon. And, certainly, electrical engineering studies electricity from an engineering perspective. But physics and electrical engineering are typically considered to be separate (albeit related) disciplines.
Should the same be said for computerscience(which would study algorithms) and computerengineering(which would study computers and, perhaps, software engineering)?
7. Arden (1980, p. 9) suggests, but does not endorse, a “committee-produced, all- purpose” definition:33 “computer science is the study of the design, analysis, and execution of algorithms, in order to better understand and extend the appli- cability of computer systems”. Note that this avoids the science-vs.-engineering quandary, by its use of ‘study’, and tries to cover all the ground. Arden sug- gests, however, that his entire book should be taken as the “elaboration” of this definition.
Isn’t this like saying that CS is what computer scientists do?
8. “Computer Scienceis the science of using computers to solve problems” (George Washington University Department of Computer Science, 2003) (see also Roberts 2006, p. 5). Because this definition doesn’t limit the kind of problems being solved, it has the advantage of illuminating the generality and in- terdisciplinarity of CS. And, because it implicitly includes the software (algorith- mic) side of computing—after all, you can’t use a computer to solve a problem unless it has been appropriately programmed—it nicely merges the computer- vs.-algorithm aspects of the possible definitions. Something more neutral could just as easily have been said: Computer science is the science of solving prob- lems computationally, or algorithmically—after all, you can’t solve a problem that way without executing its algorithmic solution on a computer.
But can there really be ascienceof problem solving? And, even if there could be, is it CS? Or is thatallthat CS is?
33That is, a “klunky” one designed to be acceptable to a variety of competing interests. The standard joke
about such definitions is that a camel is a horse designed by a committee. See http://en.wikipedia.org/wiki/ Design by committee
9. As we mentioned in§2.6.2.2, McGinn (2015b) argues that philosophy is a sci- ence just like physics (which is an empirical science) or mathematics (which is a “formal” science), likening it more to the latter than the former (p. 85). To make his argument, he offers this characterization of science:
[W]hat distinguishes a discourse as scientific are such traits as these: rigor, clarity, literalness, organization, generality (laws or general principles), tech- nicality, explicitness, public criteria of evaluation, refutability, hypothesis testing, expansion of common sense (with the possibility of undermining common sense), inaccessibility to the layman, theory construction, symbolic articulation, axiomatic formulation, learned journals, rigorous and lengthy education, professional societies, and a sense of apartness from na¨ıve opin- ion. (McGinn, 2015b, p. 86)
Does CS fit that characterization?
10. In§3.5.4, we considered the possibility that CS is not a “coherent” discipline. Consider the following interpretation of the blind-men-and-the-elephant story:
The man at the tail is sure he has found a snake; the man at the tusks believes he’s holding spears. Through teamwork, they eventually discover the truth. “But what if they were wrong?” [magician Derek] DelGaudio asks onstage. “What if that thing was some sort of magical creature that had a snake for a nose and tree-trunk legs, and they convinced it it was an elephant? Maybe that’s why you don’t see those things anymore.” (Weiner, 2017)
Might CS have been such a “magical creature”? Is it still?
11. In this chapter, we asked what CS is: Is it a science? A branch of engineer- ing? Or something else? But we could also have responded to the question with another one: Does it matter? Is it the case that, in order for a discipline to be respectable, it has to be (or claim to be!) a science? Or is it the case that a disci- pline’s usefulness is more important? (For instance, whether or not medicine is a science, perhaps what really matters is that it is a socially useful activity that draws upon scientific—and other!—sources.)34
So: Does it matter what CS is? And what would it mean for a discipline to be “useful”?
12. Wing (2016) says this about computational thinking:
I argued that the use of computational concepts, methods and tools would transform the very conduct of every discipline, profession and sector. Some- one with the ability to use computation effectively would have an edge over someone without. So, I saw a great opportunity for the computer science community to teach future generations how computer scientists think. Hence “computational thinking.”
Howdocomputer scientists think? At the very least, we might say that they thinkprocedurally. Is that the same as saying that they thinkalgorithmically? We might also say that they thinkrecursively. Because procedural theories of computability (such as Turing’s) are logically equivalent to recursive theories (such as G¨odel’s), is procedural (or algorithmic) thinking the same as recursive thinking? Is thinkingabstractlypart of computational thinking, as in the case of procedural abstraction (or is thinking abstractly merely something that is more generally part of thinking “logically” or “scientifically”)? Are there other ways in which computer scientists think that is unique to computer science?
13. A related (but distinct) question is: What is a computer scientist? Bill Gasarch
(https://blog.computationalcomplexity.org/2018/09/what-is-physicist-mathematician.html)
considers a number of reasons why the answer to this question is not straightfor- ward: Does it depend on whether the person is in a CS department? Whether the person’s degree is in CS? What the person’s research is? For example, the computer scientist Scott Aaronson received a prize in physics, yet he insists that he his not a physicist (Aaronson, 2018). Read Gasarch’s post and try to offer some answers. (We’ll return to this issue in§15.4.4.)
Chapter 4
What Is Science?
Version of 20 January 2020 DRAFT c2004–2020 by William J. Rapaport Science is the great antidote to the poison of enthusiasm and superstition. —Adam Smith (1776, V.1.203)
The most remarkable discovery made by scientists is science itself. The discov- ery must be compared in importance with the invention of cave-painting and of writing. Like these earlier human creations, science is an attempt to control our surroundings by entering into them and understanding them from inside. And like them, science has surely made a critical step in human development which cannot be reversed. We cannot conceive a future society without science.
—Jacob Bronowski (1958, my italics)
[A] science is an evolving, but never finished, interpretive system. And funda- mental to science . . . is its questioning of what it thinks it knows. . . . Scientific knowledge . . . is a system for coming to an understanding.
—Avron Barr (1985)
Science isall aboutthe fact that we don’t know everything.
Science is the learning process.
—Brian Dunning (2007)
[S]cience is not a collection of truths. It is a continuing exploration of mysteries. —Freeman Dyson (2011b, p. 10)
4.1
Readings
In doing these readings, remember that our ultimate question is whether CS is a science. 1. Required:
• Either:
– Okasha, Samir (2002),Philosophy of Science: A Very Short Introduction
(Oxford: Oxford University Press).
∗ This is my favorite introduction to philosophy of science, although it’s an entire (but short) book. You may read it instead of any of the following.
• Or:
(a) Kemeny, John G. (1959),A Philosopher Looks at Science
(Princeton: D. van Nostrand),
https://openlibrary.org/works/OL5174372W/A philosopher looks at science
– Introduction, pp. ix–xii
– Ch. 5, “The [Scientific] Method”, pp. 85–105.
– You can skim Ch. 10, “What Is Science?”, pp. 174–183, because his an- swer is just this: A science is any study that follows the scientific method. (b) Popper, Karl R. (1953), “Science: Conjectures and Refutations”, in Karl R. Popper, Conjectures and Refutations: The Growth of Scientific Knowledge
(New York: Harper & Row, 1962),
https://faculty.washington.edu/lynnhank/Popper.doc
(c) Kuhn, Thomas S. (1962),The Structure of Scientific Revolutions
(Chicago: University of Chicago Press),
Ch. IX, “The Nature and Necessity of Scientific Revolutions”,
http://www.marxists.org/reference/subject/philosophy/works/us/kuhn.htm 2. Recommended:
(a) Hempel, Carl G. (1966),Philosophy of Natural Science(Englewood Cliffs, NJ: Prentice-Hall), “Scope and Aim of this Book”,
http://www.thatmarcusfamily.org/philosophy/Course Websites/Readings/ Hempel%20-%20Philosophy%20of%20Natural%20Science.pdf
• On empirical vs. non-empirical sciences.
(b) Kolak, Daniel; Hirstein, William; Mandik, Peter; & Waskan, Jonathan (2006),Cog- nitive Science: An Introduction to Mind and Brain(New York: Routledge),
§4.4.2. “The Philosophy of Science”, http://books.ranvier.ir/download.php?file=
Cognitive%20Science%20An%20Introduction%20to%20the%20Mind%20and%20Brain.pdf (c) Papineau, David (2003), “Philosophy of Science”, in Nicholas Bunnin & E.P. Tsui-
James (eds.),The Blackwell Companion to Philosophy, 2nd edition(Malden, MA: Blackwell): 286–316,
https://svetlogike.files.wordpress.com/2014/02/the-blackwell-companion-to-philosophy-2ed-2002. pdf
• Focus on pp. 286–290 (induction, falsificationism),
& pp. 294–305 (instrumentalism, realism, theory, observation, evidence, pes- simistic meta-induction, epistemology, causation)
4.2
Introduction
All these processes are very complex, and they tend to follow the rule that the more you find out about them, the more you discover that you didn’t know . . . . That is both the joy and the frustration of science . . . .
—Gregory L. Murphy (2019,§1)
We have seen that one answer to our principal question—What is CS?1—is that it is a science (or that parts of it are science). Some say that it is a science of computers, some that it is a science of algorithms or procedures, some that it is a science of information processing. And, of course, some say that it is not a science at all, but that it is a branch of engineering. In Chapter 5, we will explore what engineering is, so that we can decide whether CS is a branch of engineering. In the present chapter, we will explore what it means to be a science, so that we can decide whether CS is one (or whether parts of it are).
In keeping with the definition of philosophy as thepersonal searchfor truth by rational means (§2.7), I won’t necessarily answer the question, “Is CS a science?”. But I will provide considerations to helpyoufindand defendan answer thatyoulike. It is more important for you to determine an answer for yourself than it is for me to present you with my view; this is part of what it means to do philosophyinthe first personfor
the first person. And it is very important for you to be able todefendyour answer; this is part of what it means to be rational (it is the view that philosophy is intimately related to critical thinking). We will follow this strategy throughout the rest of the book.