5 Does ‘life’ have a definition?

Carol E. Cleland, University of Colorado

Christopher F. Chyba, Princeton University

5.1 Introduction a position analogous to that of a sixteenth-century

investigator trying to define ‘water’ before the advent
The question ‘‘What is life?’’ is foundational to biology
of molecular theory. The best she or he could do would
and especially important to astrobiologists who may
be to define it in terms of sensible properties, such as its
one day encounter utterly alien life. But how should
being wet, transparent, odorless, tasteless, thirst
one approach this question? One widely adopted strat-
quenching, and a good solvent. But no amount of
egy among scientists is to try to define ‘life.’1 This
observational or conceptual analysis of these features
chapter critically evaluates this strategy. Drawing
will reveal that water is H2O. Yet, as we now know,
from insights gained by philosophical investigations
‘‘H2O’’ is the scientifically most informative answer to
into the nature of logic and language, we argue that it
the question ‘‘What is water?’’ Analogously, in the
is unlikely to succeed. We propose a different strategy,
absence of a general theory of the nature of living
which may prove more fruitful in searches for extra-
systems, analysis of the features that we currently asso-
terrestrial life.
ciate with life is unlikely to provide a particularly
We begin in Section 5.2 by reviewing the history of
informative answer to the question ‘‘What is life?’’
attempts to define ‘life,’ and their utility in searches for
extraterrestrial life. As will become apparent, these
definitions typically face serious counterexamples, 5.2 Attempts to define ‘life’
and may generate as many problems as they solve.
The history of attempts to define ‘life’ is very long,
To explain why attempts to define ‘life’ are fraught
going back at least to Aristotle, who defined ‘life’ in
with so many difficulties, we must first develop the
terms of the capacity to reproduce (Aristotle, De Anima
necessary philosophical background. Therefore, in
415a22–415b2; Matthews, 1977; but see also Shields,
Sections 5.3 and 5.4 we discuss the general nature of
1999). To this day, there remains no broadly accepted
definition and of so-called theoretical identity state-
definition of ‘life’ (Chyba and McDonald, 1995). The
ments. Section 5.5 then applies the material developed
scientific literature is filled with suggestions; decades
in these sections to the project of defining ‘life.’ We
ago Sagan (1970) catalogued physiological, metabolic,
argue that the idea that one can answer the question
biochemical, Darwinian (which he called ‘‘genetic’’),
‘‘What is life?’’ by defining ‘life’ is mistaken, resting
and thermodynamic definitions, along with their coun-
upon confusions about the nature of definition and its
terexamples. There have been many other attempts2
capacity to answer fundamental questions about natu-
(see, e.g., Schrödinger, 1945; Monod, 1971; Feinberg
ral categories (Cleland and Chyba, 2002).
and Shapiro, 1980; Dyson, 1985; Kamminga, 1988;
To answer the question ‘‘What is life?’’ we require
Fleischaker, 1990; Joyce, 1994, 1995; McKay, 1994;
not a definition but a general theory of the nature of
Shapiro and Feinberg, 1995; Bedau, 1996; Rizzotti
living systems. In the absence of such a theory, we are in
et al., 1996; Adami, 1998; Kauffman, 2000; Conrad
and Nealson, 2001; Harold, 2001; Schulze-Makuch
1
Single quotation marks around a word indicate that it is being et al., 2002). All typically face important problems, in
mentioned as opposed to being used. Definitions provide one exam-
ple. Another example is the claim that ‘life’ has four letters; contrast
2
this with the very different claim that life originated on Earth around Lahav (1999: 117–21) compiles 48 definitions of life (with citations)
four billion years ago. offered from 1855 to 1997.

Planets and Life: The Emerging Science of Astrobiology, eds. Woodruff T. Sullivan, III and John A. Baross. Published by Cambridge University
Press. # Cambridge University Press 2007.
120 Does ‘life’ have a definition?

that they include phenomena that most are reluctant to extraterrestrial life to date. Details of the experiments
consider alive, or exclude entities that clearly are alive performed and their results are given in Sections 18.5.2
(Chyba and McDonald, 1995). and 23.2. The basic approach was to conduct experi-
Consider a few attempted definitions by way of illus- ments with the martian soil to test for the presence of
tration (Sagan, 1970). A metabolic definition, for exam- metabolizing organisms, and indeed the results of the
ple, might be based on the ability to consume and labeled release experiment in particular were not unlike
convert energy in order to move, grow, or reproduce. what had been expected for the presence of life (Levin
But fire, and perhaps even automobiles, might be said to and Straat, 1979; Levin and Levin, 1998). But in the
satisfy some or all of these criteria. A thermodynamic end, The Viking biology team’s consensus was for a
definition might describe a living system as one that nonbiological interpretation (Klein, 1978; 1979; 1999),
takes in energy in order to create order locally, but this strongly influenced by the failure of the Viking gas
would seem to include crystals, which like fire would not chromatograph mass spectrometer (GCMS) to find
generally be considered alive. A biochemical definition any organic molecules to its limits of detection in the
would be based on the presence of certain types of soil with sample heating up to 500 8C (Biemann et al.,
biomolecules, yet one must worry that any such choice 1977). This instrument had not been intended to conduct
could in the future face exceptions in the form of systems a ‘‘life-detection’’ experiment, but de facto did so, implici-
that otherwise appear alive but are not made of our tly employing a biochemical definition. Moreover, the
particular favored molecules. Genetic or Darwinian def- GCMS would not have detected as many as !106
initions are now more generally favored than any of bacterial cells per gram of soil (Klein, 1978; Glavin et al.
these other definitions, but these too face drawbacks 2001; Bada 2001), and it now appears that oxidation of
and will be discussed in detail in Section 5.2.2 below. meteoritic organics on the martian surface may have
Another approach has been not so much to define produced nonvolatile organic compounds that would
life as simply to list its purported characteristics (e.g., not have been easily detectable (Benner et al., 2000).
Mayr, 1982; Koshland, 2002). But essentially the same Correctly interpreted or not, the result was psychologi-
difficulties arise in this approach; for example, Schulze- cally powerful: no (detected!) organics, no life. Chyba and
Makuch et al. (2002) present a list of nonbiological Phillips (2001; 2002) have presented a list of lessons to be
parallels to various supposedly distinguishing criteria learned from this experience – one lesson is that any in situ
of life such as metabolism, growth, reproduction, and search for extraterrestrial life should employ more than
adaptation to the environment. one definition of life so that results can be intercompared.
Nevertheless, the philosophical question of the def- Of course, if there were really one correct, known
inition of ‘life’ has increasing practical importance, as definition of ‘life,’ this would be an unnecessary strat-
laboratory experiments approach the synthesis of life egy. Currently, it is the Darwinian definition that seems
(as measured by the criteria of some definitions), and as most accepted. We examine this definition below, but
greater attention is focused on the search for life on shall see that rather than providing us with an unassail-
Mars (Chapter 18) and Jupiter’s moon Europa able definition, it instead presents fresh dilemmas.
(Chapter 19). In particular, definitions of ‘life’ are
often explicit or implicit in planning remote in situ
5.2.2 The Darwinian definition
searches for extraterrestrial life. The design of life-
detection experiments to be performed on Europa Darwinian (sometimes called genetic) definitions of
(e.g., Chyba and Phillips, 2001; 2002) or Mars (e.g., ‘life’ hold that life is ‘‘a system capable of evolution
Nealson and Conrad, 1999; Banfield et al., 2001; by natural selection’’ (Sagan, 1970). One working ver-
Conrad and Nealson, 2001) by spacecraft landers sion that is popular within the origins-of-life commun-
depends on decisions about what life is, and what ity is the ‘‘chemical Darwinian definition’’ (Chyba
observations will count as evidence for its detection and McDonald, 1995), according to which ‘‘life is a
(Chapters 22 and 23). This is clearly illustrated by the self-sustained chemical system capable of undergoing
story of the Viking mission’s search for life on Mars. Darwinian evolution’’ (Joyce, 1994, 1995). Joyce (1994)
explains that ‘‘the notion of Darwinian evolution sub-
sumes the processes of self-reproduction, material con-
5.2.1 Viking’s search for life on Mars
tinuity over a historical lineage, genetic variation, and
The Viking mission’s search for life on Mars in the mid- natural selection. The requirement that the system be
1970s remains the only dedicated in situ search for self-sustained refers to the fact that living systems
5.3 Definitions 121

contain all the genetic information necessary for their do not themselves evolve, so that many living entities
own constant production (i.e., metabolism).’’ The in our world are not, by the Darwinian definition,
chemical Darwinian definition excludes computer or examples of ‘‘life.’’ The Darwinian definition refers
artificial ‘‘life’’ through its demand that the system to a system that at least in some cases must contain
under consideration be ‘‘chemical’’; it also excludes more than one entity; with this reasoning Victor
biological viruses, by virtue of the ‘‘self-sustained’’ Frankenstein’s unique creation (Shelley, 1818), for
requirement. example, is not ‘‘life’’ even though it is a living entity.
Some researchers (e.g., Dawkins, 1983; Dennett, But this resolution needs to be explained as more than
1995), on the other hand, do not restrict Darwinian an ad hoc move to shave from the definition bedeviling
evolution to chemical systems, explicitly leaving open entities that we would otherwise call examples of ‘‘life,’’
the possibility of computer life. This reflects the func- but which cause trouble for a particular definition.
tionalist view (e.g., Sober, 1992) that Darwinian evolu- Finally, there is a practical drawback to Darwinian
tion is a more general process that can be abstracted definitions. In an in situ search for life on other planets,
from any particular physical realization. In this view, how long would we wait for a system to demonstrate
it is not the computer that is alive but rather the that it is ‘‘capable’’ of Darwinian evolution, and under
processes themselves. The artificial vehicle of the what conditions (Fleischaker, 1990)? This objection,
computer, produced by human beings, has a status however, is not decisive in itself, since an operational
no different from that of the artificial glassware that objection is not an objection in principle, and ways
might be used in a laboratory synthesis of organic (see Chao, 2000) might be found to operationalize the
life. It is thus not surprising that, according to this definition.
view, ‘‘living’’ systems or ecosystems can in fact be We have focused on Darwinian definitions because
created in a computer (e.g., Rasmussen, 1992; Ray, they are currently in vogue, especially in light of the
1992). great successes of the RNA world model for the origin
Yet this too may seem unsatisfactory: a computer of life (Gilbert, 1986; Chapters 6 and 8). Nevertheless,
simulation of cellular biochemistry is a simulation of as we have discussed, all of the popular versions of the
biochemistry, and not biochemistry itself. No com- Darwinian definition face similar severe challenges.
puter simulation of photosynthesis, for example, is
actually photosynthesis since it does not yield authentic
5.3 Definitions
carbohydrates; at best, it yields simulated carbohy-
drates. So why should a computer simulation of ‘‘life’’ To understand why attempts to define ‘life’ prove so
be called life itself, rather than a simulation of life? On difficult, we now develop the philosophical back-
the functionalist view, the simulation is life, because life ground for the nature of definition. Definitions are
is an abstract process independent of any particular concerned with language and concepts. For example,
physical realization. the definition ‘‘ ‘bachelor’ means unmarried human
There are further problems with Darwinian defini- male’’ does not talk about bachelors. Instead, it explains
tions, in addition to the quandary regarding computer the meaning of a word, in this case ‘bachelor,’ by dis-
‘‘life.’’ It is possible (though not generally favored secting the concept that we associate with it. As this
among current theories of the origin of life on Earth) example illustrates, every definition has two parts. The
that early cellular life on Earth or some other world definiendum is the expression being defined (‘bachelor’)
passed through a period of reproduction without and the definiens is the expression doing the defining.
DNA-type replication, during which Darwinian evolu-
tion did not yet operate (e.g., Dyson, 1985; Rode, 1999;
5.3.1 Varieties of definition
New and Pohorille, 2000; Pohorille and New, 2000). In
this hypothesis, protein-based creatures capable of Many different sorts of things are commonly called
metabolism predated the development of exact replica- ‘‘definitions.’’ In this section we will discuss only those
tion based on nucleic acids. If such entities were to be that are relevant to understanding the problem of pro-
discovered on another world, Darwinian definitons viding a scientifically useful definition of ‘life’; for more
would preclude them from being considered alive. on definitions, see, for example, Audi (1995).
There is an additional simple objection to the Lexical definitions report on the standard meanings
Darwinian definition, namely that individual sexually of terms in a natural language. Dictionary definitions
reproducing organisms in our DNA-protein world provide a familiar example. Lexical definitions contrast
122 Does ‘life’ have a definition?

with stipulative definitions, which explicitly introduce opposed to lists or gestures (the ‘dog’ example). We will
new, often technical, meanings for terms. The follow- return to this important point later.
ing stipulative definition introduces a new meaning for The most informative definitions specify the mean-
an old term: ‘work’ means the product of the magni- ings of terms by analyzing concepts and supplying a
tude of an acting force and the displacement due to its noncircular synonym for the term being defined. In
action. Stipulative definitions are also used to intro- philosophy, such definitions are known as full or com-
duce invented terms, e.g., ‘electron’ (means basic unit plete definitions. But because philosophers sometimes
of electricity), or ‘gene’ (means basic unit of heredity). use these expressions to designate more fine-grained
Unlike lexical definitions, stipulative definitions are distinctions, we shall use the term ideal definition.
arbitrary in the sense that rather than reporting on
existing meanings of terms, they explicitly introduce
5.3.2 Ideal definitions
new meanings.
Another familiar type of definition is the ostensive Ideal definitions explain the meanings of terms by
definition. Ostensive definitions specify the meaning of relating them to expressions that we already under-
a term merely by indicating a few (ideally) prototypical stand. It is thus important that the definiens make use
examples within its extension; the extension of a term is of neither the term being defined nor one of its close
the class of all the things to which it applies. An adult cognates; otherwise the definition will be circular.
who explains the meaning of the word ‘dog’ to a child Defining ‘line’ as ‘‘a linear path’’ is an example of an
by pointing to a dog and saying ‘‘that is a dog’’ is explicitly circular definition, while an implicitly circu-
providing an ostensive definition. Someone who defines lar definition is defining ‘cause’ as ‘‘something that
‘university’ as ‘‘an institution such as the University produces an effect.’’ Someone who does not under-
of Colorado, Stanford University, Universidad de stand the meaning of ‘cause’ will also not understand
Guadalajara, and Cambridge University’’ is also provid- the meaning of ‘effect’ since ‘effect’ means something
ing an ostensive definition. that is caused. Many lexical definitions suffer from the
Operational definitions provide an important related defect of circularity, which is why philosophers dislike
form of definition. Like ostensive definitions, opera- dictionary definitions.
tional definitions explain meanings via representative The definition of ‘bachelor’ (as ‘‘unmarried human
examples. They do not, however, directly indicate male’’) with which we began this discussion provides a
examples, but instead specify procedures that can be salient illustration of an ideal definition. It is not circu-
performed on something to determine whether or not it lar since the concept of being unmarried, human, and
falls into the extension of the definiendum. An example male does not presuppose an understanding of the
of an operational definition is defining ‘acid’ as ‘some- concept of bachelor. The definiens thus provides an
thing that turns litmus paper red.’ The definiens speci- informative analysis of the meaning of ‘bachelor.’ An
fies a procedure that can be used to determine whether ideal definition may thus be viewed as specifying the
an unknown substance is an acid. Operational defini- meaning of a term by reference to a logical conjunction
tions are particularly important for our discussion of properties (being unmarried, human, and male), as
since many astrobiologists, e.g., one of these authors opposed to representative examples (ostensive defini-
(Chyba and McDonald (1995), McKay (1994), Nealson tion), or a procedure for recognizing examples (opera-
and Conrad (1999), and Conrad and Nealson (2001)) tional definition). The conjunction of descriptions
have called for the use of operational definitions in determines the extension of the definiendum by specify-
searches for extraterrestrial life. The problem with ing necessary and sufficient conditions for its applica-
operational definitions is that they do not tell one tion. A necessary condition for falling into the extension
very much about what the items falling under the def- of a term is a condition in whose absence the term does
iniendum have in common. The fact that litmus paper not apply and a sufficient condition is a condition in
turns red when placed in a liquid doesn’t tell us much whose presence the term cannot fail to apply.
about the nature of acidity; it only tells us that a partic- Most purported ideal definitions face borderline
ular liquid is something called ‘acid.’ In other words, cases in which it is uncertain as to whether something
operational definitions differ from ostensive defini- satisfies the conjunction of predicates supplied by the
tions primarily in the manner in which they pick out definiens. A good example is the question of whether a
the representative examples of items falling under the ten-year-old boy is a bachelor. Moreover, even if one
definiendum, namely, indirectly by means of ‘‘tests,’’ as resolves such cases by adding additional conditions
5.4 Natural kinds and theoretical identity statements 123

(e.g., adult) to the definiens, there will always be other

borderline cases (e.g., the status of eighteen-year-old
males). Language is vague. This is brought forcefully
home by the classic example of trying to distinguish a
bald man from a man who is not bald in terms of the
number of hairs on his head. The fact that we cannot
specify a crisp boundary does not show that there is no
difference between being bald and not being bald. Ideal
definitions that specify both necessary and sufficient
conditions are rare. Nevertheless, ignoring the problem
of borderline cases, we can often construct fairly sat-
isfactory approximations. If definitions of ‘life’ faced
nothing more serious by way of counterexamples than
borderline cases (e.g., viruses), there might not be
insurmountable problems. But they have more serious
problems. Just as good definitions of ‘bachelor’ or
‘bald man’ must deal with, respectively, forty-year-old
unmarried men and men sporting thick heads of hair,
so good definitions of ‘life’ must deal with quartz crys-
tals and candle flames, which are (presumably) clearly
not alive.

5.4 Natural kinds and theoretical

identity statements
Ideal definitions specify meanings by providing a
‘‘complete’’ (within the constraints of vagueness) ana-
FIGURE 5.1. The distillation of aqua vitae, a form of ‘water.’ From
lysis of the concepts associated with terms. They work Das Buch zu Distillieren by Hieronymus Braunschweig (Strassburg,
well for terms such as ‘bachelor’ or ‘fortnight’ or 1519). (From Roberts (1994: 100); courtesy British Library.)
‘chair,’ which designate categories whose existence
depends solely on human interests and concerns.
Indeed, it is hard to imagine a better answer to the
question ‘‘What is water?’’ by defining the natural
question ‘‘What is a bachelor?’’ than ‘‘an unmarried,
kind term ‘water.’ One could try to define ‘water’ by
adult human male.’’
reference to its sensible properties, features such as
Ideal definitions do not, however, supply good
being wet, transparent, odorless, tasteless, thirst
answers to questions about the identity of natural
quenching, and a good solvent. (This is analogous to
kinds – categories carved out by nature, as opposed to
some suggested definitions of ‘life,’ e.g., that of
human interests, concerns, and conventions.3 This
Koshland (2002).) Unlike the definition of ‘bachelor,’
issue is particularly important for our purposes since
however, this definition of ‘water’ is not simply a mat-
it seems likely (but not certain) that ‘life’ is a natural
ter of linguistic convention. Nevertheless, reference to a
kind term – that whether something is living or non-
list of sensible properties cannot exclude things that
living represents an objective fact about the natural
superficially resemble water but are not in fact water.
world. Consider, for example, trying to answer the
As an example, the alchemists, impressed by water’s
powers as a solvent, identified nitric acid and mixtures
of hydrochloric acid as water, the former being known
as aqua fortis (‘‘strong water’’) and the latter as aqua
3
Some philosophers of science (known as ‘‘anti-realists’’) reject, to
regia (‘‘royal water’’); aqua vitae (‘‘water of life’’) was a
greater or lesser degrees, claims that there are knowable, mind-
independent facts, entities, or laws. We cannot engage with this
mixture of alcohols (Roberts, 1994; Fig. 5.1). Even
literature here; for an introduction see Audi (1995) and references today we commonly classify as ‘water’ various liquids
therein. that greatly differ in their sensible properties, e.g., salt
124 Does ‘life’ have a definition?

water, muddy water, and distilled water. Which of the properly speaking an element, but can be decomposed
sensible properties (e.g., transparency or tastelessness) and recombined.’’ It took more than analysis of sensi-
of the various things called ‘water’ are the important ble properties to definitively settle questions about the
ones? Five hundred years ago Leonardo da Vinci proper classification of such ostensibly different sub-
(1513) expressed this dilemma well: stances as ice and steam.
Notice that the identification of water with H2O
And so it [water] is sometimes sharp and sometimes does not have the character of an ideal definition.
strong, sometimes acid and sometimes bitter, some- It cannot be viewed as explicating the concept that
times sweet and sometimes thick or thin, sometimes has historically been associated with the term ‘water’
it is seen bringing hurt or pestilence, sometimes since that concept encompasses stuff varying widely in
health-giving, sometimes poisonous. So one would chemical and physical composition. Moreover, in
say that it suffers change into as many natures as daily discourse we still use the word ‘water’ for things
are the different places through which it passes. that are not pure H2O. The claim that water is H2O
And as the mirror changes with the colour of its began as a testable empirical conjecture (situated
object so it changes with the nature of the place within Lavoisier’s new theoretical framework for chem-
through which it passes: health-giving, noisome, istry), and it is now considered so well confirmed that
laxative, astringent, sulphurous, salt, incarnadined, most scientists characterize it as a fact. Nevertheless, it
mournful, raging, angry, red, yellow, green, black, remains a scientific hypothesis. It is conceivable (even if
blue, greasy, fat, thin. extraordinarily unlikely) that we may someday discover
that current molecular theory is wrong in some impor-
Without an understanding of the intrinsic nature of tant respect and that water is not H2O, just as Planck
water, there is no definitive answer to the question and Einstein showed a century ago that the wave theory
‘‘What is water?’’ Given an understanding of the molec- of light was incomplete and that light also behaves like a
ular structure of matter, however, such quandaries particle. If the claim that water is H2O represented an
disappear. Water is H2O – a molecule made of two ideal definition, we could not admit the possibility that
atoms of hydrogen and one atom of oxygen. H2O is water might not be H2O any more than we can conceive
what salt water, muddy water, distilled water, and even of a married bachelor or a month-long fortnight.
acidic solutions have in common, despite their obvious It is sometimes claimed that theoretical identity
sensible differences. The identification of water with statements such as ‘‘water is H2O,’’ ‘‘temperature is
H2O explains why liquids (e.g., nitric acid) that (in mean kinetic energy,’’ and ‘‘sound is a compression
some ostensibly important ways) resemble water are wave’’ represent stipulative definitions. On this view
not water; their molecular composition is more than they amount to nothing more than linguistic decisions
H2O alone. Furthermore, the identification explains to take familiar terms from common language and give
the behavior of what we call ‘water’ under a wide them wholly new technical meanings within the context
variety of chemical and physical circumstances. The of a currently accepted theory (Nagel, 1961). The prima
identification holds regardless of whether the water is facie problem with this account is that it prevents us
in any of its familiar solid, liquid or vapor phases, and it from making sense of the idea that these statements tell
will hold equally well in less familiar high-pressure us something new about the stuff designated by the old
solid phases. Indeed, before the advent of modern familiar terms (‘water,’ ‘temperature,’ ‘sound’). Rather
chemistry, it was not widely recognized that ice, than learning something new, in this view we are merely
steam, and liquid water are phases of the same kind attaching new concepts (identifying descriptions) to old
of stuff. Some ancient Greeks (for example, terms, and hence only changing the way we talk about
Anaximenes) believed that steam was a form of ‘‘air’’ the world. One might be tempted to say that this is the
(Lloyd, 1982: 22). As late as the late seventeenth cen- way language works: if one changes the concept asso-
tury, ice and water were thought to be different ‘‘spe- ciated with a word radically enough, then one is no
cies.’’ The Aristotlelian view of water as one of the four longer talking about the same thing. However, such an
basic elements out of which all matter is constructed approach, associated with the philosopher John Locke
only began to fall into disfavor in the late eighteenth (and exploited by Thomas Kuhn in his famous argu-
century with work such as Antoine Lavoisier’s paper ments for the incommensurability of scientific theo-
(1783) entitled ‘‘On the nature of water and on experi- ries), faces serious logical problems; we discuss these
ments that appear to prove that this substance is not in detail in Appendices 5.1–5.2 at the end of this
5.5 What is ‘life’? 125

chapter. For this reason most contemporary philoso- scientific theory can be conclusive, but someday we
phers reject the view that theoretical identity state- may have a well-confirmed, adequately general theory
ments are stipulative definitions. Some radical of life that will allow us to formulate a theoretical
changes in the concept of an old word are the result identity statement providing a scientifically satisfying
of discovering that we were wrong about the familiar answer to the question ‘‘What is life?’’
phenomenon that the word designates; for more detail,
see Appendix 5.3. Put more concretely, we know some-
5.5.1 Dreams of a general theory of life
thing about water that Aristotle and Anaximenes
didn’t know: water is not a primitive element, but a In order to formulate a convincing theoretical identity
molecular compound. statement for life we need a general theory of living
systems. The problem is that we are currently limited to
only one sample of life, namely, terrestrial life.
5.5 What is ‘life’?
Although the morphological diversity of terrestrial
Let us return to the definition of ‘life.’ If (as seems life is enormous, all known life on Earth is extraordi-
likely, but not certain) life is a natural kind, then narily similar in its biochemistry. With the exception of
attempts to define ‘life’ are fundamentally misguided. some viruses, the hereditary material of all known life
Definitions serve only to explain the concepts that we on Earth is DNA of the same right-handed chirality.
currently associate with terms. As human mental enti- Furthermore, life on Earth utilizes 20 amino acids to
ties, concepts cannot reveal the objective underlying construct proteins, and these amino acids are typically
nature (or lack thereof) of the categories designated of left-handed chirality. These biochemical similarities
by natural kind terms. Yet when we use a natural lead to the conclusion that life on Earth had a single
kind term, it is this underlying nature (not the concepts origin. Darwinian evolution then explains how this
in our heads) that we are interested in. ‘Water’ means common biochemical framework yielded such an
whatever the stuff in streams, lakes, oceans, and every- amazing diversity of life. But because the biochemical
thing else that is water has in common. We currently similarities of all life on Earth can be explained in terms
believe that this stuff is H2O, and our belief is based on of a single origin, it is difficult to decide which features
a well-confirmed, general scientific theory of matter. of terrestrial life are common to all life, wherever it may
We cannot, of course, be absolutely positive that be found. Many biochemical features that currently
molecular theory is the final word on the nature of strike us as important (because all terrestrial life shares
matter; conclusive proof is just not possible in science. them) may derive from mere chemical or physical con-
Nevertheless, our current scientific concept of water as tingencies present at the time life originated on Earth
H2O represents a vast improvement over earlier con- (Sagan, 1974). In the absence of a general theory of
cepts based on superficial sensory experience. If we living systems, how can we discriminate the contingent
someday discover that molecular theory is wrong, we from the essential? It is a bit like trying to come up with
will change the concept that we associate with ‘water,’ a theory of mammals when one can observe only
but we will still be talking about the same thing. zebras. What features of zebras should one focus
Analogously to ‘water,’ ‘life’ means whatever cya- upon – their stripes, common to all, or their mammary
nobacteria, hyperthermophilic archaeobacteria, amoe- glands, characteristic only of the females? In fact, the
bae, mushrooms, palm trees, sea turtles, elephants, mammary glands, although present in only some
humans, and everything else that is alive (on Earth or zebras, tell us more about what it means to be a mam-
elsewhere) has in common. No purported definition of mal than do the ubiquitous stripes. Without access to
‘life’ can provide a scientifically satisfying answer to the living things having a different historical origin, it is
question ‘‘What is life?’’ because no mere analysis using difficult and perhaps ultimately impossible to formu-
human concepts can reveal the nature of a world that late an adequately general theory of the nature of living
lies beyond them. The best we can do is to construct systems.
and empirically test scientific theories about the general This problem is not unique to life. It reflects a simple
nature of living systems, theories that settle our classi- logical point. One cannot generalize from a single
ficatory dilemmas by explaining puzzling cases – why example. What makes the case of life seem different is
things that are alive sometimes lack features that we the amazing diversity of life on Earth today. We risk
associate with life and why things that are non-living being tricked into thinking that terrestrial life provides
sometimes have features that we associate with life. No us with a variety of different examples. But biochemical
126 Does ‘life’ have a definition?

analyses coupled with knowledge of evolution reveals non-living is really living, or vice versa.5 But to be in a
that much of this diversity is a historical accident. Had position to formulate such a theory will require a wider
the history of the Earth been different, life on Earth diversity of examples of life. Current laboratory inves-
today would certainly be different. ‘‘How different?’’ is tigations (e.g., research on the hypothesized prebiotic
a crucial question for astrobiology. In the absence of a ‘‘RNA World’’ on Earth) and empirical searches for
general theory of living systems, one simply cannot extraterrestrial life are important steps in supplying
decide. In essence, the common origin of contemporary these examples. Until the formulation of such a theory,
terrestrial life blinds us to the possibilities for life in we will not know whether such a theoretical identity
general. statement for life exists.
A look at some popular definitions of ‘life’ illus-
trates the problem of trying to identify the nature of
5.5.2 How to search for extraterrestrial life
life in the absence of an adequately general theoretical
framework for living things. Many definitions There remains the problem of how to hunt for extra-
(e.g., Conrad and Nealson, 2001; Koshland, 2002) terrestrial life without either a definition of ‘life’ or a
cite sensible properties of terrestrial life – features general theory of living systems. One approach is to
such as metabolism, reproduction, complex hierarchi- treat the features that we currently use to recognize
cal structure, and self-regulation. But defining ‘life’ in terrestrial life as tentative criteria for life (as opposed
terms of sensible properties is analogous to defining to defining criteria). These features will then necessarily
‘water’ as being wet, transparent, tasteless, odorless, be inconclusive; their absence cannot be taken as suffi-
thirst quenching, and a good solvent. As we have dis- cient for concluding that something is not alive.
cussed, reference to sensible properties is unable both Therefore they cannot be viewed as providing opera-
to exclude things that are not water (e.g., nitric acid) tional definitions of ‘life’ (in the strict sense of that
and to include everything that is water (e.g., ice). term). The purpose of using tentative criteria is not to
Similarly, this approach will be unsuccessful for defin- definitively settle the issue of whether something is
ing ‘life.’4 alive, but rather to focus attention on possible candi-
Definitions of ‘life’ that do not make reference to dates, namely, physical systems whose status as living
sensible properties typically suffer from being too gen- or non-living is genuinely unclear. Accordingly, the
eral. Definitions of ‘life’ based on thermodynamics criteria should include a wide diversity of the fea-
provide good examples. As discussed in Section 5.2, it tures of terrestrial life. Indeed, diversity is absolutely
is difficult to exclude systems (e.g., crystals) that are crucial (Cleland, 2001; 2002) when one is looking for
clearly non-living without introducing ad hoc devices evidence of long past extraterrestrial life, e.g., in the
(Chyba and McDonald, 1995). Similarly, the ‘‘chemical martian meteorite ALH84001 (Section 18.5.3), or with
Darwinian’’ definition discussed earlier (Section 5.2.2) instrument packages delivered to ancient martian flood
excludes problematic cases (such as artificial or com- plains or to europan frozen ice ‘‘ponds’’ (Section 19.7).
puter life) by simply stipulating that something must be Some features for shaping searches for extraterrestrial
a chemical system in order to qualify as living. If we had life (whether extant or extinct) may not even be univer-
an adequate theoretical framework for understanding sal to terrestrial life. For example, features that are
life, we could avoid the problem of being too general common only to life found in certain terrestrial envi-
without resorting to ad hoc devices. ronments may prove more useful for searching for life
New scientific theories change old classifications, in analogous extraterrestrial environments than fea-
for example by uniting mass and energy under mass– tures that are universal to terrestrial life. Similarly,
energy, or, less profoundly, by splitting jade into the features that are uncommon or non-existent among
two minerals jadeite and nephrite. A general theory of non-living terrestrial systems may make good criteria
living systems might well change our current classifica- for present or past life, even if they are not universal
tions of living and non-living. These changes in classi- to living systems, because they stand out against a
fication will be convincing only if an empirically tested, background of non-living processes. The chains of
general theory of living systems can explain, chemically pure, single-domain magnetite crystals
for example, why a system that we once viewed as found in ALH84001 provide a potential example

4 5
For further discussion of the relation between the concept of life and Other possibilities include three distinct categories of life, or no
the features that we use to recognize it, see Lange (1996). distinct categories, but rather a continuum.
 Appendix 5.1 Locke’s theory of meaning 127

(Section 18.5.3). If (as is still quite controversial) it extraordinarily diverse, so diverse that it seems highly
turns out that these chains can only be produced bio- improbable that any conjunction of descriptions could
genically (except perhaps under circumstances that are distinguish everything that is a game from everything
exceedingly unlikely to occur in nature), then they will that is not a game. Wittgenstein concludes that there
provide a good biosignature for life, despite the fact are no necessary and sufficient conditions for being a
that most terrestrial bacteria do not produce them. game. According to Wittgenstein, what distinguishes
The basic idea behind our strategy for searching for games from things that are not games is family resem-
extraterrestrial life is to employ empirically well- blance: if an item has enough of the pertinent proper-
founded, albeit provisional, criteria that increase the ties, then it is a game. But whether concepts are
probability of recognizing extraterrestrial life while identified with clusters or conjunctions of descriptions,
minimizing the chances of being misled by inadequate the question of whether an item falls into the extension
definitions. This is similar in spirit (though with greater of a term is taken to be completely settled by whether it
care given to the limitations of ‘definition’) to sugges- fits the descriptions that we happen to associate with
tions that in situ searches for extraterrestrial life should the term. The upshot is that anything that fits our
rely when possible on contrasting definitions of life current concept of water qualifies as ‘‘water.’’ If our
(Chyba and Phillips, 2001; 2002). Unlike efforts that concept of water were completely founded on sensible
focus on a favored definition, our suggestions are per- properties and the sensible properties that we deemed
haps closest to the strategy proposed by Nealson and to be most important failed to exclude nitric acid, then
his colleagues, who (despite their liberal use of the word not only would we call nitric acid ‘‘water’’ (which,
‘‘definition’’) emphasize the use of a number of widely historically speaking, we once did), but on the
diverse biosignatures (atmospheric, hydrospheric, and Lockean view, nitric acid would actually be water. On
lithospheric) (Conrad and Nealson, 2001; Storrie- this view, there is no possibility of discovering that we
Lombardi et al. 2001; Chapter 23). The important are wrong – that our descriptions are too inclusive or
point, however, is that our strategy is deliberately exclusive – since the only thing that qualifies an item as
designed to probe the boundaries of our current con- a member of the extension of a term is whether it
cept of life. It is only in this way that we can move happens to fit the descriptions that we associate with
beyond our Earth-centric ideas and recognize genu- the term. If we change our concept of water by stipulat-
inely weird extraterrestrial life, should we be fortunate ing, in the context of a new theory, that water is H2O,
enough to encounter it. And it is only by keeping the then we are no longer talking about the same thing.
boundaries of our concept of life adaptable and open to Thus Aristotle, who held that water is an indivisible
unanticipated possibilities that we can accrue the element, cannot be interpreted as talking about the
empirical evidence required for formulating a truly same thing that we are talking about when we use the
general theory of living systems. word ‘water’ because, for us, water is a composite of
hydrogen and oxygen atoms.
Locke’s theory is unable to distinguish natural kind
Appendix 5.1 Locke’s theory
terms from non-natural kind terms. Locke was fully
of meaning
aware of this; his solution was to bite the bullet, and
The idea that theoretical identity statements represent reject the distinction. In a revealing discussion Locke
stipulative definitions receives support from a prob- (1689: Book III, Chapter XI, Section 7) argues that the
lematic theory of meaning associated with seventeenth- seventeenth-century debate over whether bats are birds
century philosopher John Locke (1689; see Schwartz, has little scientific merit since the (seventeenth-century)
1977, for a review). According to this theory, the mean- concepts of bat and bird are compatible with either
ing of any term in a language is completely exhausted position; for Locke, the debate is merely verbal. Yet
by the concepts associated with it, and concepts are in hindsight this seems wrong. The question of whether
identified with descriptions. On some versions of the bats are birds is not merely verbal – a matter of what
theory, concepts are analyzed as clusters (rather than description we decide to associate with ‘bat’ and ‘bird.’
logical conjunctions) of descriptions. Wittgenstein’s Indeed, we have discovered that the things we call
oft-cited analysis of the meaning of the word ‘game’ ‘‘bats’’ are far more like mammals than birds. It is
provides a good illustration (see Wittgenstein, 1953; instructive to compare this situation with an analogous
also Schwartz, 1977). The items (e.g., chess, solitaire, argument over whether bachelors could be married. No
water polo, charades) that we call ‘‘games’’ are one can discover that bachelors are married. Any one
128 Does ‘life’ have a definition?

who claims that they have done so either does not whether construed as conjunctions or as clusters of
understand the meaning of ‘bachelor’ or, alternatively, descriptions.
is simply stipulating (vs. discovering) a new meaning There are a number of different versions of the
for ‘bachelor.’ In other words, unlike the debate over new theory of meaning. All of them, however, agree
whether bats are birds, the question of whether that meaning involves reference, and reference is not
bachelors can be married is purely verbal. An determined by concepts. The word ‘water’ means
adequate theory of meaning should be able to explain whatever has the same intrinsic nature as the stuff
the difference between common nouns like ‘bat’ and that we typically call ‘water’ regardless of the descrip-
‘bachelor.’ tions that we happen to associate with it. While it is
undeniable that we use descriptions (derived from our
sensible experiences with paradigmatic examples) to
Appendix 5.2 John Locke and
recognize things as water, these descriptions do not
Thomas Kuhn
(as in the old Lockean view) determine what it is for
The Lockean view of meaning underlies Thomas something to be water. Thus something can fit descrip-
Kuhn’s famous argument for the incommensurability tions that we associate with ‘water’ and yet fail to
of scientific theories (Kuhn, 1962). When the defining qualify as water by virtue of having the wrong intrinsic
descriptions associated with a term drastically change, nature.
as happens in scientific revolutions, the Lockean theory This point is illustrated by Hilary Putnam (1973;
says that the meaning of the term also drastically 1975), a founder of the new theory, in a well-known
changes. Thus the term ‘mass’ means something dras- thought experiment. Putnam asks us to suppose that
tically different in Newtonian mechanics (where mass is there existed a fantastic planet called ‘‘Twin Earth.’’
conserved) than it does in the special theory of relativ- Twin Earth is like Earth, but the liquid called ‘‘water’’
ity (where only mass–energy is conserved). The upshot on Twin Earth is not H2O but a different liquid whose
is that we can’t say that the special theory of relativity chemical formula is abbreviated as ‘‘XYZ.’’ XYZ and
tells us something new about the thing referred to by H2O have the same sensible properties; XYZ is wet,
the old term ‘mass.’ Rather than expanding our transparent, odorless, tasteless, and a good solvent.
knowledge of the natural world, on Kuhn’s account, In Putnam’s thought experiment, Twin Earthers from
new scientific theories only alter our conceptual frame- the seventeenth century (before molecular theory
work. Yet this conclusion seems wrong. Surely we appeared) and seventeenth-century Earthlings have
know more about the natural world than we did a the same concept of water. A seventeenth-century
hundred years ago! Earthling might well believe that there is water on
The inadequacy of the Lockean framework for Twin Earth. But that conclusion would be wrong.
meaning cannot, in our view, be overstated. A success- The stuff on Twin Earth that looks like water is not
ful theory of meaning must account for indisputable water because it is not H2O, even though Twin Earthers
facts about language and thought; after all, language and Earthlings might not understand this until the end
and thought are the subject matter of a theory of mean- of the eighteenth century.
ing. It is undeniable that we speak and think differently It is important to understand the point of Putnam’s
about natural kinds than we do about conventional thought experiment. The fact that it makes little scien-
kinds. Because it treats the meaning of every term as tific sense to speak of Twin Earth being just like Earth
just a matter of convention – as depending only upon except for the chemical composition of water is not
the concepts that we happen to associate with it – the relevant to his argument. Putnam is making a point
Lockean view cannot accommodate this difference; it about language and concepts. Language is used to
lacks the resources to explain it. describe many kinds of situations, from actual to hypo-
thetical (e.g., what if Al Gore had been the US
President in 2003?), to fantastic (e.g., the adventures
Appendix 5.3 A new theory of meaning of the young wizard Harry Potter). An adequate theory
In contemporary philosophy, the Lockean view has of meaning must do justice to hypothetical and fantas-
been challenged by a new theory of meaning tic situations as well as factual ones. Putnam’s thought
(Schwartz, 1977 gives a review). This new theory solves experiment about Twin Earth demonstrates that the
the problems of the old theory by dispensing with the meaning of a natural kind term is not fully captured
whole project of identifying meanings with concepts, by the descriptions that we associate with it. If it were,
References 129

we would have to conclude that our seventeenth cen- change the subject and began talking about something
tury Earthling is correct about there being water on entirely different. We learn something new about an old
Twin Earth. familiar subject, and this may include discovering that
We now have the tools to evaluate the proposal that our language and concepts have badly misled us.
theoretical identity statements (the theoretical identity
statement ‘‘water is H2O’’) are stipulative definitions.
On either the old or the new theory of meaning, defini- R EFERENCES
tions are concerned only with language and concepts. If Adami, C. (1998). Introduction to Artificial Life. New York:
statements such as ‘‘water is H2O’’ are stipulative defini- Springer-Verlag.
tions, then (à la Kuhn) they don’t tell us anything new Aristotle. De Anima. Trans. by J. A. Smith (1941). In The
about the world of nature. They represent nothing more Basic Works of Aristotle, ed. R. McKeon. New York:
than linguistic decisions to attach concepts (H2O) Random House.
derived from theoretical frameworks (molecular theory) Audi, R. (ed.) (1995). The Cambridge Dictionary of
to familiar old terms (‘water’). But this interpretation Philosophy. Cambridge: Cambridge University Press.
does not do justice to the way we think and speak
Bada, J. (2001). State-of-the-art instruments for detecting
about theoretical identity statements: we take them to extraterrestrial life. Proc. Natl. Acad. Sci. USA, 98, 797–800.
be making defeasible (capable of being invalidated)
Banfield, J. F., Moreau, J. W., Chan, C. S., Welch, S. A., and
claims about the old familiar world of experience. On
Little, B. (2001). Mineralogical biosignatures and the search
the problematic old theory of meaning, this aspect of our for life on Mars. Astrobiology, 1(4), 447–465.
conceptual structure and linguistic behavior could not be
Bauer, M. (1968). Precious Stones Vol. II. New York: Dover
explained. The upshot was that scientific debates over the
Publications.
underlying nature of natural kinds had to be interpreted
Bedau, M. (1996). The nature of life. In The Philosophy of
(e.g., as in Locke’s analysis of whether bats are birds) as
Artificial Life, ed. Margaret Boden. Oxford: Oxford
merely verbal. The new theory of meaning restores the
University Press, 332–357.
connection between language and the world, and allows
Benner, S., Devine, K., Matueeva, L., and Powell, D. (2000).
us to make good sense of our intuitions about the con-
The missing molecules on Mars. Proc. Natl. Acad. Sci. USA,
tingent empirical status of theoretical identities.
97, 2425–2430.
Rather than viewing theoretical identity statements
Biemann, K., Oro, J., Toulmin, P. III, et al. (1977). The search
as stipulative definitions, it is more accurate to construe
for organic substances and inorganic volatile compounds in
them as empirical conjectures, situated within the con-
the surface of Mars. J. Geophys. Res., 82, 4641–4658.
text of a well-confirmed scientific theory, about a cat-
Chao, L. (2000). The meaning of life. BioScience, 50, 245–250.
egory of items treated in common discourse as a natural
kind. Theoretical identities are contingent in the sense Chyba, C. F. and McDonald, G. D. (1995). The origin of life
that (unlikely as it now seems in some cases) we might in the solar system: current issues. Ann. Rev. Earth Planet. Sci.,
23, 215–249.
someday discover that they are false. Moreover, it is
important to keep in mind that there is no guarantee Chyba, C. F. and Phillips, C. B. (2001). Possible ecosystems
and the search for life on Europa. Proc. Natl. Acad. Sci. USA,
that our best scientific theories will carve up the world in
98, 801–804.
exactly the same way as natural language. As an exam-
ple, jadeite and nephrite were once included under the Chyba, C. F. and Phillips, C. B. (2002). Europa as an abode of
common term ‘jade,’ but it is now clear from chemical life. Orig. Life Evol. Biosph., 32, 47–68.
analysis and microscopic examination that they are dif- Cleland, C. E. (2001). Historical science, experimental science,
ferent (Bauer, 1968; Putnam, 1975); the term ‘jade’ does and the scientific method. Geology, 29, 978–90.
not designate a (single) natural kind after all. Similarly, Cleland, C. E. (2002). Methodological and epistemic differ-
in the context of the right theoretical framework, we ences between historical science and experimental science.
may discover that what we thought were different Philosophy of Science, 69, 474–496.
natural kinds are actually part of the same natural Cleland, C. E. and Chyba, C. F. (2002). Defining ‘life’. Orig.
kind. To cite another example from mineralogy, we Life Evol. Biosph., 32, 387–393.
now know that rubies and sapphires, despite their strik- Conrad, P. G. and Nealson, K. H. (2001). A non-Earth-centric
ing sensible differences, are members of the natural kind approach to life detection. Astrobiology, 1, 15–24.
corundum (Al2O3) (Bauer, 1968). In short, when an old Da Vinci, Leonardo (1513). Quoted in Whitcombe, C., Leonardo
theory is replaced by a new theory, we do not simply da Vinci and Water, at witcombe.sbc.edu/water/artleonardo.html.
130 Does ‘life’ have a definition?

Trans. by MacCurdy (2003: 734) from Il Codice Arundel, No. 263, (trans. Carmen Giunta. Available online at webserver.lemoy-
fol. 57r. ne.edu/faculty/giunta/laveau.html).
Dawkins, R. (1983). Universal Darwinism. In Evolution from Levin, G. V. and Levin, R. L. (1998). Liquid water and life on
Molecules to Men, ed. D. S. Bendall. Cambridge: Cambridge Mars. Proc. SPIE – The International Society for Optical
University Press, 403–425. Engineering, 3441, 30–41.
Dennett, D. C. (1995). Darwin’s Dangerous Idea. New York: Levin, G. V. and Straat, P. A. (1979). Completion of the
Simon and Schuster. Viking labeled release experiment on Mars. J. Mol. Evol., 14,
Dyson, F. (1985). Origins of Life. Cambridge: Cambridge 167–183.
University Press. Lloyd, G. E. R. (1982). Early Greek Science: Thales to
Feinberg, G. and Shapiro, R. (1980). Life Beyond Earth: Aristotle. London: Chatto & Windus.
Intelligent Earthlings, Guide to the Universe. New York: Locke, J. (1689). An Essay Concerning Human Understanding.
William Morrow. Oxford: Oxford University Press.
Fleischaker, G. R. (1990). Origins of life: an operational def- MacCurdy, Edward (2003). The Notebooks of Leonardo da
inition. Orig. Life Evol. Biosph., 20, 127–137. Vinci, Definitive Edition in One Volume. Old Saybrook:
Gilbert, W. (1986). The RNA world. Nature, 319, 618. Konecky and Konecky.

Glavin, D., Schubert, M., Botta, O., Kminek, G., and Bada, J. Matthews, G. B. (1977). Consciousness and life. Philosophy,
(2001). Detecting pyrolysis products from bacteria on Mars. 52, 13–26.
Earth Planet. Sci. Lett., 185, 1–5. Mayr, E. (1982), The Growth of Biological Thought.
Harold, F. M. (2001). Postscript to Schrodinger: so what is Cambridge, MA: Belknap Press.
life? Am. Soc. Microbio. News, 67, 611–616. McKay, C. P. (1994). Origins of Life. In Van Nostrand Reinhold
Joyce, G. F. (1994). Forword. In Origins of Life: the Central Encyclopedia of Planetary Sciences and Astrogeology, eds.
Concepts, eds. Deamer, D. and Fleischaker, G., pp. xi–xii. Shirley, J. and Fairbridge, R. New York: Van Nostrand.
Boston: Jones & Bartlett. Monod, J. (1971). Chance and Necessity: an Essay on the Natural
Joyce, G. F. (1995). The RNA world: life before DNA Philosophy of Modern Biology. London: Alfred A. Knopf.
and protein. In: Extraterrestrials – Where Are They? II. Nagel, E. (1961). The Structure of Science: Problems in the
eds. Zuckerman, B. and Hart, M., 139–151. Cambridge: Logic of Scientific Explanation. New York: Harcourt, Brace &
Cambridge University Press. World.
Kamminga, H. (1988). Historical perspective: the problem of Nealson, K. H. and Conrad, P. G. (1999). Life: past, present
the origin of life in the context of developments in biology. and future. Philos. Trans. R. Soc. Lond. B Biol. Sci., 354,
Orig. Life Evol. Biosph., 18, 1–11. 1923–1939.
Kauffman, S. (2000). Investigations. Oxford: Oxford New, M. and Pohorille, A. (2000). An inherited efficiencies
University Press. model of non-genomic evolution. Simulation Practice and
Klein, H. P. (1978). The Viking biological experiments on Theory, 8, 99–108.
Mars. Icarus, 34, 666–674. Pohorille, A. and New, M. (2000). Models of protocellular struc-
Klein, H. P. (1979). Simulation of the Viking biology experi- tures, functions, and evolution. In Frontiers of Life, eds. Palyi, G.,
ments: an overview. J. Mol. Evol., 14, 161–165. Zucchi, C., and Caglioti, L. New York: Elsevier, 37–42.

Klein, H. P. (1999). Did Viking discover life on Mars? Orig. Putnam, H. (1973). Meaning and reference. J. Philos., 70,
Life Evol. Biosph., 29, 625–631. 699–711.

Koshland, D. E. (2002). The seven pillars of life. Science, 295, Putnam, H. (1975). The meaning of meaning. In: Mind,
2215–2216. Language and Reality: Philosophical Papers, Volume 2.
Cambridge: Cambridge University Press, 215–271.
Kuhn, Thomas S. (1962). The Structure of Scientific
Revolutions. Chicago: The University of Chicago Press. Rasmussen, S. (1992). Aspects of information, life, reality, and
physics. Artificial Life, 2, 767–74.
Lahav, N. (1999). Biogenesis: Theories of Life’s Origins. New
York: Oxford University Press. Ray, T. S. (1992). An approach to the synthesis of life.
Artificial Life, 2, 371–408.
Lange, M. (1996). Life, ‘‘artificial life,’’ and scientific explan-
ation. Philosophy of Science, 63, 225–244. Roberts, G. (1994). The Mirror of Alchemy: Alchemical Ideas
in Images, Manuscripts and Books. Toronto: University of
Lavoisier, A. L. (1783). On the nature of water and on experi-
Toronto Press.
ments which appear to prove that this substance is not strictly
speaking an element but that it is susceptible of decomposition Rode, B. M. (1999). Peptides and the origin of life. Peptides,
and recomposition. Observations sur la Physique, 23, 452–455 20, 773–786.
Further reading and surfing 131

Rizzotti, M. (ed.) (1996). Defining Life. Padova: Padova Sober, E. (1992). Learning from functionalism – prospects for
University Press. strong artificial life. Artificial Life, 2, 749–765.
Sagan, C. (1970). Life. In: Encyclopaedia Britannica (fifteenth Storrie-Lombardi, M., Hug, W., McDonald, G., Tsapin, A.,
edn.), 22, 985–1002. and Nealson, K. (2001). Hollow cathode ion laser for deep
Sagan, C. (1974). The origin of life in a cosmic context. Orig. ultraviolet Raman spectroscopy and fluorescence imaging.
Life Evol. Biosph., 5, 497–505. Rev. Sci. Instrum., 72, 4452–4459.

Schrödinger, E. (1945). What is Life? The Physical Aspect of Wittgenstein, L. (1953). Philosophical Investigations. Trans. by
the Living Cell. Cambridge: Cambridge University Press. G. E. M. Anscombe. New York: Macmillan.

Schulze-Makuch, D., Guan, H., Irwin, L., and Vega, E.

(2002). Redefining life: an ecological, thermodynamic, and F URTHER READING AND SURFING
bioinformatic approach. In Fundamentals of Life, eds.
Palyi, G., Zucchi, C., and Caglioti, L. New York: Elsevier, Boyd, R., Gasper, P. and Trout, J. D. (1997). The Philosophy
of Science. Cambridge, MA: MIT Press.
169–179.
Kim, J. and Sosa, E. (1995). A Companion to Metaphysics.
Schwartz, S. P. (1977). Introduction. In Naming, Necessity,
and Natural Kinds, ed. Schwartz, S. Ithaca: Cornell Oxford: Blackwell Publications.
University Press, 13–41. Newton-Smith, W. H. (2001). A Companion to the Philosophy
Shapiro, R. and Feinberg, G. (1995). Possible forms of life of Science. Oxford: Blackwell Publications.
in environments very different from the Earth. In: Rosenberg, A. (2000). Philosophy of Science. New York:
Extraterrestrials: Where Are They? eds. Zuckerman, B. and Routledge.
Hart, M. Cambridge: Cambridge University Press, 165–172. Schwartz, S. P. (ed.) (1977). Naming, Necessity, and Natural
Shelley, M. (1818). Frankenstein: or the Modern Prometheus. Kinds. Ithaca, NY: Cornell University Press.
London: Lackington, Hughe Harding, Mavor & Jones. Stanford Encyclopedia of Philosophy at plato.stanford.edu.
Shields, C. (1999). Order in Multiplicity. Oxford: Oxford A marvellous and authoritative source about all aspects of
University Press. philosophy and philosophers.