Life, what exactly is it?

Mark A. Bedua (Author)

Song Ci and Fan Xingchen (translators)

Fascinating life

The Earth's surface is covered with life and is often easily recognizable. Cats, carrots, bacteria are all alive, bridges, soap bubbles, sand grains are dead. But it's no secret that biologists don't have a precise definition of life. Because biology is the science of life, one might expect that the exploration of the question of the nature of life will occupy an important place in contemporary biology and the philosophy of biology. In fact, however, biologists and philosophers today do not discuss the nature of life. Many people believe that the definition of life is not directly related to current biological research (Sober, 1992; Taylor, 1992). When biologists discuss life in general, they usually marginalize their discussions and give something specious. But now everything has changed.

Today, the question of the nature of life is a hot topic. The economic industries that manipulate life are growing rapidly. Biotechnologies such as genetic engineering, cloning, and high-speed DNA sequencing give us new, unprecedented power to reinvent life. A recent development is that we are able to use synthetic genomics to redesign life to our specifications (Gibbs, 2004; Brent, 2004). In this area, Craig Venter has attracted attention by touting the use of commercially available artificial cells to clean the environment or produce alternative fuels (Zimmer, 2003). The current competition for "wet" artificial life through the synthesis of the smallest artificial cells or protocells in test tubes (Szostak, Bartel & Luisi, 2001; Rasmussen et al., 2004; Luisi, 2006; Rasmussen et al., 2007) also focuses on the essentials of life. Because this race requires a consensus on a definition of life that transcends life forms as we know it. The social and ethical impact of creating protocells also requires a deeper understanding of the nature of life. Currently on the origin of life (Oparin, 1964; Crick, 1981; Shapiro, 1986;Eigen, 1992; Morowitz, 1992; Dyson, 1999; Luisi, 1998) and the debate over intelligent design (Pennock, 2001) are even more in full swing.

In addition, a recent development called "soft" artificial life also focuses on the nature of life, which attempts to develop software systems to give them the essential properties of life (Bedau, 2003a). Soft artificial life creates very realistic software systems that some people think are really alive (Langton, 1989a; Ray, 1992), but others think that the whole idea that computer simulations are life is ridiculous (Patte, 1989).

In addition, recent achievements have been made in "hard" artificial life, such as the first widely used commercial home vacuum cleaner robot Roomba (Brooks, 2002), and a walking robot designed and manufactured by improving automated rapid prototyping (Lipson & Pollack, 2000). The existence of these robots inevitably leads to the question of whether a device made of plastic, silicon and steel can really be called alive? These scientific developments have increased the uncertainty of how to precisely divide organisms.

Biology generalizes the forms that life can take, but this generalization is also based on real life forms. Biologists have studied a large number of different model organisms, such as E. coli (a common bacterium), Caenorhabditis elegans (a nematode), and Drosophila melabini (a type of drosophila). The selection of as many model organisms as possible can best reflect the forms that life can take, and make the most comprehensive generalization of life on Earth. But all life on Earth is on Earth. Therefore, these generalizations about life currently depend on the sample size. Maynard Smith (1998) once pointed out that artificial life can help alleviate this problem. Life in nature has an astonishing diversity. But they are only a small fraction of all possible life forms. Any time we can use software, hardware, or wetware to synthesize a system that demonstrates the core properties of life, there are plenty of opportunities to expand our empirical understanding of what life is.

Three great men in the history of philosophy have advanced ideas about life, and these views still resonate in contemporary discussions. In On the Soul, Aristotle argues that life is a kind of ruler nested by various faculties such as metabolism, sensation, and movement. This nested ruler of power corresponds to Aristotle's conception of "soul" or spiritual power, so Aristotle essentially associates life with the mind. As part of a comprehensive replacement of Aristotle's philosophy and science, Descartes believed that life was merely a complex but purely mechanical process of operation performed by a material machine. Descartes believed that life was fundamentally different from the mind, which he considered to be a mode of consciousness. Descartes outlined the details of his mechanical hypothesis about life in his Treatise on Man. A few generations later, Kant sought to combine Descartes' materialist views with the unique autonomy and purpose of life in his Critique of Judgment.

Understanding the nature of life is not simple. It requires an investigation into something real and extremely complex, and these things have enormous potential for creativity and power to change the face of the planet (Margulis & Sagan, 1995). This survey is necessarily interdisciplinary, and it will examine the astonishing number of perspectives on life. For example, wholeness, internal stability, purposefulness, and evolutionability, these interesting and subtle properties are seen as characterizing life. But a precise definition of life remains elusive, in part because of critical edges such as viruses and spores, and more recently the creation of artificial life. More complex is the centrality of life in a series of philosophical conundrums that in turn involve important philosophical questions such as emergence, computation, and the mind. It is therefore expected that views on life are diverse. Some apply familiar philosophical theories such as functionalism. Others use biochemical or genetic explanations and mechanisms. Still others emphasize processes such as metabolism and evolution. The diversity of perspectives on life is interesting in itself and worth elaborating on.

Life phenomena

Life has many characteristics and marginal cases, showing many puzzles. The remainder of this chapter is devoted to explaining these phenomena.

A compelling fact of life is the characteristics and unique signs it exhibits. We usually think that these characteristics are neither sufficient nor necessary for life. Still, they are typical features of life. Different people give lists of different life characteristics; for example, Maynard Smith, 1986; Farmer & Belin, 1992; Mayr, 1997; Gánti, 2000。 But most of the features in the list are basically overlapping. Another point worth noting is that there are also clear differences in the features listed on the list. A good example is the feature given by Gánti (or what he calls "standard").

Gánti's characteristics fall into two categories: real (or absolute) and latent. The true standard of life prescribes the full and necessary conditions for the existence of the individual living organism as life. The "real" standard of life proposed by Gánti in 2003 is as follows:

(1) Integrity

An organism is a separate entity that cannot be subdivided without losing its basic properties. If the parts of an organism are separated and no longer interact, it cannot be considered alive.

(2) Metabolism

Individual organisms absorb matter and energy from their surroundings and cause them to undergo chemical reactions. Seeds lack an active metabolism when they are dormant, but they can be said to be alive when external conditions reactivate their metabolism. For this reason, Gánti distinguishes general things into four states: alive, dormant, dead, or impossible to live.

(3) Intrinsic stability

When the living environment is constantly changing, the organism is able to maintain a stable state of internal processes. By changing and adapting to the dynamic external environment, the organism maintains its overall structure and organization. This involves detecting changes in the environment and compensating for the internal changes they cause, thereby acting as a protector of the entire internal organization.

(4) Effective information portability system

A living system must be able to store the information that is used for its growth and functioning. Because this information can be copied, it is obtained by reproduction so that the offspring inherit this information. Errors in the transmission of information can "mutate" this genetic information, while natural selection can screen for the resulting genetic variation.

(5) Flexible control

Processes in the organism are controllable, which allows organisms to persist and flourish. This control involves a flexibility to adapt that can often be improved as experiences increase.

Corresponding to these "real" standards of life, Gánti also proposes "potential" standards of life. A single living organism cannot demonstrate the underlying standards of life. The decisive feature of the potential standard of life is that if there are enough organisms to demonstrate the potential standard, then life can reproduce and sustain itself on a planet. In this regard, Gánti proposed the following three points:

(1) Growth and reproduction

Older animals and sterile plants and animals are living beings, but none of them are able to reproduce. Thus, for a living organism, the ability to reproduce is neither necessary nor sufficient. But since individual organisms die, the group can survive and prosper only if certain organisms in the group reproduce. In this sense, growth and reproduction are what Gánti calls "potential" standards of life, not "real" standards of life.

(2) Evolvability

"A living system must have the capacity for genetic change, but also the capacity to evolve, that is, to produce increasingly complex and differentiated forms over a long series of successive generations" (Gánti, 2003, pp.79). Since it is not individual organisms that evolve over time, but the populations to which these organisms belong, we should say that living systems are members of populations that have the ability to evolve. Which biological population has the ability to generate greater complexity and differentiation? This remains an open question.

(3) The inevitability of death

Living systems die. This is true even for organisms that reproduce asexually through cloning, and both living individuals and their clones face death. There is no immortal system that does not die, so death is an attribute of living creatures.

Gánti's list of life standards and other people's life standards always reflects and expresses some preconceptions about life. This still doesn't seem to answer the question of what life is. The list of each life feature is constructed by someone to draw certain instances in or out by certain criteria. But where does this criterion come from, and how can we be sure that it is correct? How can we be convinced that (in line with any of the criteria on the list) reveal the nature of life? Thus it can be seen that the list of life features does not give a final answer to what life is. As our understanding of life deepens, so do our ideas. Therefore, our list of characteristics of life should also go further.

Another interesting feature of life is that there are marginal cases that lie somewhere between the living and the inanimate. Common examples are viruses and prions, which are able to replicate and spread on their own even without independent metabolism. Dormant seeds or spores are another marginal case, the most extreme of which is that of bacteria or insects that have been frozen. There are also cases where they are clearly not alive, but they still have the characteristics of living systems. No one would think of the flame of a candle as life, but its constituent molecules can retain their shape under changing conditions, a process somewhat like metabolism (Maynard Smith, 1986). Tiny clusters of clay crystals that can grow and proliferate are another marginal case, especially since they can be naturally selected in the right environment (Bedau, 1991). A race that is spreading ("proliferating"?). Forest fires are also a case of edge, where at the edge of a fire, the fire is transmitted from one tree to another, as if a bacterial population were growing at the edge. There are further marginal cases of superorganisms made up of a group of organic organisms, such as fully social insect communities, which function like a single organism. Although controversial, biologists believe that superorganic beings themselves see themselves as living beings. Another marginal case includes soft artificial life creations such as Tierra. Tierra is software that creates a spontaneously evolving group of computer programs that multiply, mutate, and evolve in computer memory. The inventors of Tierra believe that Tierra is real and alive (Ray, 1992). This will completely subvert most of us's general concept of life. The last type of edge case is the complex adaptive systems around us, such as financial markets or the World Wide Web. These cases present a large number of characteristics of life, so some argue that the simplest and most uniform explanation for the full range of life phenomena is to regard these natural complex adaptive systems as real life (Bedau, 1996, 1998).

The mystery of life

The third point is that there are many mysteries in life itself. We will list the following six points. Any description of life should explain the origin of these puzzles; more importantly, these descriptions should solve these puzzles. Some puzzles may simply be out of confusion, but others are basic and fascinating unanswered questions in nature.

(1) Origins

How is life born out of inanimate? How did biology evolve from pure chemistry? Consider a system that only undergoes chemical reactions, in which chemical reactions continuously change the concentration of chemicals, what is the difference between such a system and a system that contains life? Where is the boundary between the phenomena of life and the mere physico-chemical phenomena? How can the boundaries between the two be naturally crossed in principle and practice? Dennett argues that Darwin's interpretation solves this problem. The scheme is achieved by introducing "finite degeneration, in which the extraordinary properties sought (i.e., life) are obtained through minor, even imperceptible corrections or additions" (1995, pp. 200).

(2) Emergence

How did life come about? Attribute B depends on attribute A, and after spawning independently of attribute A, with autonomy, attribute B emerges from attribute A. Different types of dependencies and autonomous independence will produce different degrees of emergence (Bedau, 2003b). If the phenomenon of emergence involves a top-down causal force that is in principle indistinguishable, it is called a "strong" emergence. The perceptual or perceptual nature of the philosophy of mind is an example (Kim, 1999). If A and B are simultaneous, then the emergence of B from A is synchronic. This involves which attributes exist at a given moment. These properties may be changing, but the relationship between properties A and B at some point is a static snapshot of the dynamic process of these properties changing. Conversely, if attribute A precedes attribute B, and attribute B arises after attribute A is generated, then the appearance of attribute B from attribute A is dynamic. Life is a dynamic form of "weak" emergent paradigms, and unless you look at the processes by which they arise, or observe their simulations, those macroscopic properties involved are unpredictable or insignificant (Bedau, 1997, 2003b).

(3) Hierarchy

The appearance of life has a variety of structural layers. Each organism forms a hierarchy relative to the organic organization within it. The relative complexity that exists between the different kinds of organisms is organized, and another hierarchy is formed. The simplest organic tissues are prokaryotic cells, whose composition is relatively simple. More complex are eukaryotic cells containing complex organelles and nuclei. Multicellular organisms are more complex; their constituent units (individual cells) are also independent living individuals (e.g., they can maintain biological activity on their own). In addition, mammals have complex internal organs (such as the heart) that can be picked up and kept biologically active when one mammal dies, and then surgically implanted into another mammal. Two problems arise here. First, why does life tend to arise and encompass such a hierarchy? This question applies both to the complex hierarchy that all organisms collectively form, as well as to the hierarchies within each organism's internal tissues. With regard to the latter, we can ask the next question. Organisms are typically living paradigms, but we also call organs and individual cells living. For example, apoptosis is an important process in the programmed death of living cells in living organisms, and hospitals strive to make certain organs survive after death so that they can be transplanted into other people's bodies. This raises the question of whether mammals, its hearts, and the cells that make them up are alive in the same sense.

(4) Continuity

Is there a degree of life? Is life a black-and-white Boolean property, or a continuous property with a transitional gray area? Common sense favors Boolean's view: rabbits are alive, stones are not, and the discussion ends here. But there are also some edge cases like viruses, where replication cannot be achieved without a host. Spores or frozen bacteria can remain dormant and remain unchanged for a long time, but they are revived again when conditions become favorable. So are viruses and spores life in the full sense of the word? Further, when primitive life emerged from the chemical soup of former organisms, they were almost no different from their inanimate predecessors. Some have concluded that life is somewhat continuous (e.g., Cairns-Smith, 1985; Emmeche, 1994; Dennett, 1995）。 Another option is to accept the obvious distinction between life and inanimate, but allow for a small step to cross it. There are four distinctions between the two: (i) something inanimate and never alive, (ii) something that is alive, (iii) something that is dead but has ever lived, or (iv) something that is dormant but can be biologically active again. These differentiation schemes help explain the existence of some boundary cases and reclassify them (e.g., when seeds and spores are dormant without reflecting life activity). But this does not completely solve the mystery of the continuity of life phenomena, because there are also marginal cases in these four distinctions, such as the situation between life and death.

(5) Powerful artificial life

Artificial life based on software and hardware raises the question: Can our computer creations really be alive? (Langton, 1989a; Patte, 1989; Sober, 1992; Emmeche, 1992; Olson, 1997) On the one hand, certain unique carbon-based macromolecules play a vital role in the survival of all known life; on the other hand, many human beings seem to assume that life can be implemented in a properly programmed computer. There are two issues that need to be distinguished here. The first is the philosophically controversial question of why computers or robots are alive. If this problem is solved, the technical question we will face is whether it is possible to make it truly alive in this sense by building a software system or hardware device (such as a robot). The challenge here is whether we can really understand how to achieve the process of life with the right raw materials. This software-based "strong" artificial life firmly believes that the examples of artificial life software are really living things. People have similar tough stances about "hard" artificial life based on hardware devices and "wet" artificial life based on laboratory equipment. These tough positions contrast with those of undisputed "weakness" and lead to the belief that living systems can be further understood through computer models, hardware structures, and wet lab creations. Moreover, a strong version of wet artificial life intuitively seems reasonable; we are usually willing to accept that what is synthesized from scratch in the laboratory is real life. Therefore, the debate about strong artificial life mainly focuses on soft artificial life and hard artificial life.

(6) Mind

Another puzzle is whether there is an intrinsic connection between life and mind. For example, plants, bacteria, insects, and mammals have various sensitivities to the environment that affect their behavior in various ways, as well as various forms of communication between them (e.g., Dennett, 1997). These are all forms of intelligent behavior, and the relative maturity of these "mental" abilities seems to correspond to the relative maturity of these life forms and gives an explanation for them. Therefore, people naturally ask whether there is some deep connection between life and mind. Sure, the evolutionary process gives the genealogical connection between life and mind, but if what Beer says is "an act of adaptation, a ... The ability to cope with the complex, dynamic, unpredictable world in which we live, which, in fact, is the basis of [intelligence itself]", will deepen the connection between life and mind (Beer, 1990, pp.11; See also Maturana & Varela, 1987; Godfrey-Smith, 1994; Clark, 1997）。 Since all forms of life must respond in one way or another to a complex, dynamic, and unpredictable world, perhaps this adaptive flexibility binds life and mind together.

An explanation of life

People have made all sorts of attempts to describe the universal characteristics of all life. In this section, I will discuss several major explanations of life, pointing out their motivations, strengths, and weaknesses. I will also point out some skeptical positions that find these explanations useless.

Consider first the skeptical position that the nature of life has nothing to do with biology (Sober, 1992; Taylor, 1992）。 The reason for this skepticism is that biologists can continue their biological research regardless of whether life can be adequately defined, and whichever view of life ultimately wins. It must be acknowledged, however, that recent developments, such as attempts to make the smallest artificial cells from scratch, do require scientists to begin to determine what is the essence of life, even if there is no precise definition of these ideas. Thus, even if it was once considered irrelevant, it is no longer the case. For a man can begin to construct a minimum form of life only when he has at least feasible assumptions about the minimum sufficient conditions of life. Otherwise he doesn't know what to do.

The second form of skepticism is the belief that there are no necessary and sufficient conditions for life to be identified, but consists only of a set of things with the similarity of the Wittgenstinian family. Different forms of life may have different properties or characteristics, but there are exceptions to the nature of individual individuals in each group. These properties are usually possessed by living organisms, but they are not strictly necessary or sufficient. Farmer and Belin listed eight traits: processes; self-reproduction; self-characterized information storage; metabolism; functional interaction with the environment; interdependence of parts; stability under perturbation; and the evolutionary ability of members of the population. Then, they explain, when they were desperate to try to find something more precise than this list of features, the cluster conception of life was born.

There doesn't seem to be a single attribute that can mark life. Any attributes we give to life are either too broad to be so broad that many inanimate systems also have such characteristics, or so specific that we can find counterexamples that do not satisfy such specific features, but intuitively judge that they are alive. (Farmer & Belin, 1992, pp. 818; see Taylor, 1992)

The concept of clustering is equivalent to doubting the possibility of a unified theory of life.

One advantage of the clustering concept is that it provides a natural explanation for boundary cases. All clustering concepts inevitably have boundary cases. One of the characteristics of the cluster concept is that it cannot explain why life forms are unified by one set of features rather than another. The cluster view must accept only the given features and then use those features to identify the clusters. For this reason, this view can only determine the characteristics of life after the fact; it cannot predict or explain these characteristics. Those who think that the characteristics of life should be explained will find that the concept of clusters is not satisfactory to them.

A similar skeptical view questions the idea that life is natural. Keller (2002) says that life is an artificial concept, not a natural one, that is, it is we who give the distinction between life and non-life, not nature. This can explain those marginal cases. Since the concept of life changes with the advancement of science and technology, one should expect its boundaries to change, resulting in marginal cases. This view also provides some general rebuttals to the mysteries of life, because it is expected that a fickle artificial object will naturally produce those mysteries. Keller's argument that life is an artificial concept suggests that the current hypothesis that life has an essence only emerged 200 years ago, that the exploration of the essence of life is driven by attempts to create life from the inanimate (which tend to break the line between life and inanimate), and that new concepts resulting from scientific and technological advances violate old taxonomies, such as the living/inanimate distinction (Keller, 2002).

All of these arguments are problematic. First, all modern scientific concepts, such as matter and energy, arose at some point in human history and have evolved ever since. Thus there is no accidental, traceable, recent origin that suggests that a class is artificial, unless it is done at once in all scientific concepts. Second, building bridges from abiotic to biological in the lab does not require the elimination of the lines between the two, just as building the first airplane does not eliminate the distinction between flying and non-flying. It is important to remember that we are looking for the essence of life, not just the current concept of life.

Now, for the question "What is life?" The answer to this question is simply to give a taxonomic system of organisms. Solving this problem is tantamount to asking for an exhaustive list of all living life on Earth. It's an interesting historical question, but full of contingencies. This classification system will not be able to give information about beings that may exist but have not been found to be present. This illustrates the chauvinism of the taxonomic view, which assumes that life as we know it has exhausted all life forms. Unrelated life forms that exist outside of Earth like Europa are not listed in all such classification systems. In any case, we should also adapt our classification system at any time as technology advances, which is how we learn.

Some people give a biochemical definition of life. Taking into account the general constraints of physics and chemistry, they attempted to specify the biochemical properties that any form of life must possess (Pace, 2001; Benner, Ricardo & Carrigan, 2004). This includes thermodynamic limits, energy limits, material limits, and even geographical limits. The characteristics of life in biochemical definitions are sometimes referred to as biochemical "commonalities" of life, which are always premised on a priori descriptions of life; it states the physical, chemical, and biological possibilities of any biochemical system to satisfy previous descriptions of life. Pace (2001) and Benner et al. (2004) gave biochemical definitions of life based on evolution, so Pace and Benner focused on biochemical commonalities of genetic abilities and emphasized that molecules such as DNA can store and transmit information between generations. Biochemical definitions are often short-sighted and assume that all possible life forms are similar to familiar life forms. One can imagine starting with a different concept of life than evolution-based, such as metabolism, to finally emphasizing biochemical commonalities that differ from DNA genetic information (such as enabling open systems to maintain their structure in the case of the second law of thermodynamics).

One genetic example of the definition of biochemical life is Venter defining life as the smallest genome sufficient to sustain its existence (Hutchison et al., 1999). This view inherits the limitations of the biochemical definition. Genomes define the simplest known genomes that contain enough to sustain life. But it doesn't include genes in every life form, because the same basic life functions can be achieved through different genes. Many will question the limitations of molecular definitions on genetic traits, because the core of life involves much more than genes (Cho et al., 1999).

Scientists who make artificial cells or "protocells" from scratch acknowledge that the nature of life is controversial, but almost all have a common goal of building an independent system that can be metabolized and evolved (e.g., Rasmussen et al., 2004). That is, any chemical system that chemically integrates the following three processes at a chemical level can be considered an artificial cell. The first is the process of assembling some kind of container( such as a lipid vesicle) and surviving in it. This is followed by the repair and regeneration of the vessel and its internal inclusions, which allow the entire system to maintain metabolic processes. These chemical processes are formed and guided by a third chemical process that involves encoding information about the system ("genes") and storing it inside the system; errors ("mutations") can occur when this information is copied, so the system can evolve through natural selection. This trinitarian perspective of life requires that the three chemical processes of accommodating, metabolism, and evolution support and promote each other, thus forming a functional feedback between the three. This view of procellular life as a complete functional triad treats any biochemical implementation that fully encompasses all three processes as true life.

The philosophy of mind of the previous generation has always been dominated by functionalism: they believe that the mind is a specific input-output device, that possessing the mind is simply having a set of internal states that have causal interactions (or "executive functions") with each other, causal interactions (or "executive functions") with inputs to the environment, and outputs to the environment in a specific way. Functionalism holds a view of life that is a network implementation that performs a particular interaction process. Some processes (e.g., information processing, metabolism, purposeful activities) run through the life cycle of an organism; others (e.g., self-propagation and adaptive evolution) processes run through many generations of organisms. These processes are always achieved through some specific underlying substance, but as long as the form of the process is retained, it does not really matter which underlying substance the process is implemented by. For these reasons, functionalism is an attractive explanation for life. Chris Langton's defense of artificial life is a functionalist classic statement about life:

Life is an attribute of form, not of matter, the result of the organization of matter, not something inherent in matter itself. （Langton, 1989a，pp.41）

An important point is that a well-organized set of artificial primitives plays the same functional role as biomolecules in natural living systems, undertaking the same process of "living" as natural organisms. Thus, artificial life is true life—it is only made up of different matters compared to life that evolved on Earth. （Langton，1989a，pp.33）

We may not be sure of the details that define the course of life, and we may wish to preserve the judgment of whether the creation of artificial life really exists. However, it is difficult to deny Langton's view that the characteristic processes of life, such as metabolism, information processing, and self-reproduction, can be achieved within a broad and potentially open material range. Thus, the prospects for a certain functionalism of life seem brighter.

The main challenge of functionalism in terms of the mind is related to consciousness and the sensory quality (qualia). It is worth noting that functionalism about life does not face any similar problems. Another challenge of functionalism in terms of the mind is to explain how people's mental states are meaningful or have semantic content. Darwin's theory of natural selection gives a naturalistic explanation for many of the biological functions fulfilled by the structure of evolved life forms. This biological function gives a meaning or semantic content to the internal state of the organism, so we can say that an organism is trying to find food in order to obtain nutrients. Many philosophers are optimistic that the question of the meaning of spirit in functionalism will be solved by a Darwinian explanation of the biological function of mental states (e.g., Dennett, 1995).

Another obvious threat to life-related functionalism is that, in some relevant sense, the processes involved in life are immutable or non-computational (e.g., Emmeche, 1992). Bedau (1999) argues that the non-computational nature of life on the surface is explainable. Otherwise equal, the dominant trait produced by mutation tends to persist and spread through populations. In addition, otherwise equal, trait repetition rates in the population will change in such a tendency manner, which is a way of generally benefiting the population in a changing external environment. Dynamic patterns of repetition rates of these traits emerge as statistical patterns from microscopic-level contingencies such as natural selection, mutation, and drift. Bedau believes that these models tend to have a special flexibility. The pattern of repetition rates for this trait is not a precise, universal summary without exception, but holds only if most other conditions are the same. In addition, there are exceptions to those pattern rules, which in a way "proves that they (these pattern rules) are nothing more than by-products given to achieve some deeper adaptation goal. For example, Bedau describes a system in which mutation rates can evolve and shows that this rate of mutation that tends to evolve will keep the gene pool of the population on a "disordered edge"; but there are exceptions to this regularity, due to the deeper laws of evolution that operate mutation rates are (through mutation rates) optimally balancing the "record of (ontology)" and "creation (generated for adaptation to the environment)" in evolution (see Bedau, 1999 for details). In this case, the flexible rule reflects a potential ability (of life existence) to be able to respond appropriately to changes in an open environment. Although life can be simulated in an appropriate computer mode, this still demonstrates an immutability of the life process.

Functionalism fails to answer exactly what role these processes play in the functional representation of life. Based on Schrödinger's influence, the defining process of life is to counter the second law of thermodynamics through metabolism:

When to say that something is alive? In contrast to an inanimate thing, living matter can continue to "do something," move, exchange materials with its surroundings, and so on for longer than we expect. An organism appears so mysterious precisely because it avoids rapid decay to a "balanced" state of inactiveness: how does a living organism avoid this decay? The obvious answer is: by eating, drinking, breathing and (in the case of plants) digestion. The technical term is metabolism. （Schr dinger, 1969，pp.74-6）

The metabolism-centered view of life has attracted many people (Margulis & Sagan, 1995; Boden, 1999）。 They are closely related to the focus on self-generated ideas (Varela, Maturana, & Uribe, 1974; Maturana & Varela, 1987）。

The idea that metabolism is the core process of life has some obvious advantages, such as which explains why we intuitively think that crystals are not alive (only some molecular metabolic flow exists at the edge of the crystal, and does not exist inside the crystal). In addition, the fact that it is necessary to fight increased entropy through metabolism means that metabolism is at least necessary for all physical life forms. Metabolism also naturally explains the difference between inanimate, alive, dead, and dormant. In principle, inanimate beings cannot be metabolized, while those living beings are metabolizing. Those who died had lived and metabolized, and are now rotting. Dormant ones who were once alive but are not currently metabolizing can be metabolized again if the environment is right.

The main disadvantage of having metabolism as a comprehensive description of life is that many metabolic entities do not intuitively seem to be alive, or are not associated with life in any way. Typical examples include candle flames, swirls, and convective cells (Maynard Smith, 1986; Bagley & Farmer, 1992). These examples by themselves do not conclusively prove that metabolism is inadequate in defining life, since intuitive judgments that precede theory may be wrong. The question is whether balanced metabolism adequately explains the characteristics of life and solves the mystery of life.

Some argue that the central feature of all life is an open-ended adaptive evolutionary process. The core idea is that what can distinguish life is its ability to open up, (within a certain range) to automatically and appropriately adapt to unpredictable changes in the environment. From this perspective, life is unique in that it evolves adaptively to survive and thrive through new, intelligent strategies as the external environment changes. Maynard Smith (1975, pp. 96f; see also Mayr, 1982; Cairns-Smith, 1985) succinctly explains the legitimacy of the idea that the key to life depends on the evolutionary process of adaptation:

We will consider any physical population with reproductive, genetic and mutated properties to be alive. The legitimacy of this definition is as follows: any population with these traits will evolve through natural selection in order to better adapt to its environment. If time is sufficient, natural selection can produce any degree of adaptive complexity.

These comments illustrate how the process of adaptive evolution explains the characteristics of life, marginal cases, and mysteries (see Bedau, 1998).

There are some typical criticisms of the evolution-centric view. One counterexample is that as living but infertile organisms (e.g., mules, old people, etc.) cannot participate in the evolutionary process. The typical response is to require that organisms be produced through evolutionary processes, but not that they necessarily influence further evolution. Another so-called counterexample is an apparently inanimate system, such as the population of clay microcrystallines or a free-market economy, which evolved through natural selection. Some argue that we should accept these counterintuitive examples because evolution-centered perspectives offer such convincing explanations for the characteristics, marginal cases, and mysteries of life (e.g., Bedau, 1998).

Not all positions are opposed to each other, and many are in agreement with each other. For example, functionalism is consistent with the interpretation of the minimum life of the protocellular integration triplet. In addition, the descriptions of the nature of life all involve the biochemical characteristics of life, and many descriptions of life overlap. The problem with understanding life is determining which of these statements is true.

Understand the problems that life faces

How should we compare and evaluate descriptions of the nature of life? A straightforward answer is to see how well each theory explains the phenomena of life. It's the equivalent of doing three things: explaining the characteristics of life, explaining the edge case, and solving the mystery of life. The problem of understanding life is the problem of explaining these three things.

One of the initial difficulties was to confuse the crux of the matter. Some studies argue that the key test for any explanation of life is to conform it to our pre-theoretical intuition that what is alive and what is not (e.g., Boden, 1999). But there is something that should be asked as to why we emphasize this intuition. A good theory of life may lead us to redefine and classify life. This may change our attitudes about where life exists. Thus, while our pre-theoretical intuitions carry a certain weight, they are not inviolable.

It is also possible to ask what the word "life" means in english today. But stereotypes associated with the word "life" are commonplace and reflect our current lowest common impression of the situation of life. Therefore, it is unlikely that we can rely on the meaning of the word "life" to understand life.

Nor can we learn much by analyzing the concept of the word life. Like the meaning of "life," our current concept of life will reflect our current understanding of life. If we want to understand the true nature of phenomena with life features, marginal cases, and life mysteries, we should study natural phenomena themselves, not our words or concepts. We should expect our understanding of the phenomena of life to evolve and advance.

The interpretation of the phenomena of life involves at least a rough view of the source or essence of life, and perhaps a rough definition of life. Scientific essentialism, derived from Kripke (1980), is the philosophical view that the essence of natural classes such as water and gold is the potential causal force discovered by empirical science (see Bealer, 1987). The essence of substances such as water and gold is the chemical composition beneath their surface. Life, on the other hand, is a flexible process, not a fixed chemical. Thus, unlike water or gold, the essence of life may be contained in a network of features that explain its unique causal capacities (such as metabolism, reproduction, and sensational) processes. In this respect, life is more like heat, a specific process in this substance (polymer kinetic energy). A specific temperature (such as 23 ° C) is a specific process that can occur in all substances. Life is also a process that can occur in different kinds of matter, but unlike temperature, not all kinds of matter can be alive. In summary, mapping biochemical constraints to the kinds of matter that can instantiate life gives rise to the biochemical definition of life. It is worth noting that even if contemporary science does not agree on what life is, the scientific essentialism of life may be correct. Scientific essentialism is a philosophical view of the method of discovery of the essence of life, not a view of what the essence of life is. Further scientific progress may need to be waited to give a scientific essentialist definition of life in detail.

It is unclear whether living things have the characteristics that make them essentially life. For example, in Dennett's view, the distinction between life and non-life is a matter of degree, and life is too "interesting" to seek its essence (1995, pp. 201). In fact, contemporary biology and bio philosophy fully embrace Darwinian anti-essentialism, i.e., species have no essence, and its members have unnecessary and sufficient attributes. In contrast, the similarities between species members are only statistically significant. Species are nothing more than a cloud or a tuft of grass in an abstract space of possible features. Although some sub-regions of the possible feature space are not occupied due to incompatibility, it is a fortuitous event that exactly which acceptable sub-regions are occupied. No sub-region is more essential than any other; there is nothing special about each sub-region in the face of a fixed and unchanging Platonic essence. This anti-essentialist generalization may help explain why so many philosophers are drawn to the concept of clusters of life, which seems to be a direct result of anti-essentialism.

Darwin's anti-essentialism was directed at a narrow conception of essence that promoted the condition of necessity and sufficiency without exception and excluded marginal cases. The marginal case is one of the signs of life, so the essence of life must be broad and flexible enough to include the marginal case. One can accept Darwin's anti-essentialism while still accepting the scientific essentialism of life. In this view, the "essence" of life will be any process that explains the phenomena of life, including the characteristics of life, marginal cases, and mysteries. Life is not defined by no exceptions to limitations, but by experience. Unfortunately, contemporary philosophical terminology obscures the compatibility of Darwin's anti-essentialism and scientific essentialism in the question of life.

Clelland and Chyba (2002) argue that it is too early to give a normative definition of life because our current understanding of life is too limited. They conclude that we should wait until scientists can distinguish between a wider variety of life forms before giving a normative definition of life. However, we may now be at a time when we need to construct tentative and testable assumptions about the phenomena of life. These assumptions are most likely wrong, but they can help us find better theories (Wimsatt, 1987). When we have a good theory of life at hand, we can distill the definition of life implicit in them. Therefore, the pursuit of the definition of life is more like the pursuit of the essence of life.

Life is one of the most fundamental and complex phenomena in nature. Therefore, the explanation of life is both rich and interesting, with a complex structure. These explanations come in many forms, such as skepticism, detailed biochemical and molecular descriptions, and abstract functionalism, and they emphasize fundamental biological processes such as metabolism and evolution. The criteria for evaluating these explanations include their ability to interpret the characteristics of life, marginal cases, and the ability to solve life's mysteries. Many of the main explanations for life still lack substantial development and careful evaluation of many of these aspects. Therefore, understanding life remains an open question.

原文：Bedau, M. (2008). What is Life?https://www.cambridge.org/core/books/abs/the-nature-of-life/what-is-life-selections/295D60C5FE17B13DA3293FDB02A5370D