Sixteen Days
The Ontology of Fetal Development (1)
Barry Smith and Berit Brogaard
draft XXII
SUNY at Buffalo, September 1, 1999
Abstract
We believe that, with a combination of detailed biology and rigorous ontological analysis, it is possible to resolve the issue of when the human individual begins to exist. We lay down a set of necessary and sufficient conditions for being a human individual and we determine when these conditions are first satisfied. We conclude that the pertinent categorial change takes place around sixteen days after fertilization, and we consider a range of arguments for and against this conclusion.
0. Introduction
1. The Marks of Substance
2. The Marks of Relatively Isolated Causal Systems
3. The Hierarchical Structure of the Human Organism
4. When Does the Human Being Begin to Exist?
5. The Varieties of Substance Formation
6. The Development of the Foster
7. Alternative Thresholds
8. Twinning
9. The Concept of Niche
10. Is the Foster Connected to the Mother?
11. Is the Foster a Part of the Mother?
12. Stages in the Formation of the Human Substance
13. The Necessary and Sufficient Conditions for Being a Human Being:
An Excursus on Siamese Twins
14. How to Refute Our Argument
0. Introduction
We shall help ourselves in what follows to the Danish term 'foster', which refers in neutral fashion to the human zygote, embryo, or fetus at the different stages of its development, and we shall focus on the question of the transtemporal identity or non-identity of the foster with the human individual as it exists after birth. In order to answer this question we will need to take empirical details into account. It matters, for example, whether or not the cells constituting the early embryo are a mere collection (in much the same sense as a heap of grains of sand is a mere collection) or whether they form some sort of genuine (causal) unity. However, we shall also need to take into account the principles of ontology. This is because, while it is sufficient to use our unaided common sense in order to establish that, for example, mature adults are human beings, our common-sense concepts are in need of careful elucidation if we are to be in a position to use them to answer the much more difficult question concerning when human beings begin to exist. The concepts we shall need for this purpose - concepts of unity, identity, boundary, part, whole, dependence - are ontological concepts.
Our strategy will be to use such ontological concepts to lay down sufficient conditions for being a human being. We shall then, by examining the biological details of fetal development, establish when these conditions are first satisfied. Along the way we shall test the robustness of our theory inter alia by considering a range of possible counterexamples, entities which are not human beings but which might be advanced as satisfying the conditions listed. Unaided common sense, we might say, is a reliable tool in relation to the standard average cases which fall within the interior of any given domain. When we move out to consider examples on the outer fringes of the domain, this helps us to test the adequacy of our elucidations of such concepts.
1. The Marks of Substance
We take the ontological core of the common-sense concept of human being to be captured by the philosopher's concept of an individual physical thing or substance. (There is a related, though separate, issue of non-physical substances, to which we shall return briefly in the sequel.)
The marks of the concept of substance are as follows: (2)
1. A substance is a bearer of change. Substances undergo processes, they participate in events, and they manifest contrary qualities at different times. (John is sometimes warmer, sometimes cooler.)
2. A substance is an entity which persists self-identically through time. It is an identity through change. A substance remains, from the beginning to the end of its existence, numerically one and the same even when undergoing alterations of a range of different sorts. (John is the same substance as he was this morning, even though his temperature has changed.)
3. Each substance is extended in space, and thus it has spatial parts. (The spatial parts of John are, for example, his arms and legs, his cells and molecules.) (3)
4. Each substance possesses its own complete, connected external boundary, analogous to the boundary of a sphere or torus, which divides its interior from its exterior and at the same time separates it spatially from other substances. Substances are distinguished, through this separation, from the undetached parts of substances. (John is distinguished from his arms and legs by the fact that he possesses his own complete, connected external boundary.)
5. The boundary of a substance serves also to unify together its parts. A substance is distinguished, through the connectedness of its boundary, from heaps or aggregates of substances. (Consider the jazz band in which John plays of an evening.)
6. Each substance is an independent entity in the sense that it does not require the existence of any other specific entity in order to exist. (4) Headaches, waves and smiles are like substances in that they persist through time, but they are unlike substances because they depend for their existence on specific other things (and ultimately on substances) as their bearers or carriers.
The concept of substance is a categorial concept. This means that it is a highest-level concept obtained by generalization from lower-level concepts such as: rabbit, mammal, animal. Other categorial concepts are: event or process (terms we here use interchangeably), relation, quality, and so on. Categorial concepts are such that, if an entity is an instance of a given category at any stage of its existence, then it instantiates this same category at every stage of its existence. This means, inter alia, that a substance cannot be transformed into a process (as a cardinal number cannot be transformed into a peach or a ray of light). It is characteristic of categorial concepts that they are instantiated either fully or not at all. There are no degrees of substancehood (as there are no degrees of identity or existence). Every categorial change is therefore instantaneous.
2. The Marks of Relatively Isolated Causal Systems
The above criteria are not, as yet, sufficient for our purposes. For they are satisfied not only by human beings and other organisms but also by bounded chunks of dead matter (for example by planets, bricks or bowling balls). They are even satisfied by the whole consisting of a closed box together with various wooden blocks which have been placed inside it. And they are satisfied by a whole consisting of various large and small fish together with a sealed aquarium inside which they are swimming. Each single cell is a substance in something like this latter sense: it has a membrane surrounding a nucleus, mitochondria, endoplasmic reticulum, Golgi apparatus, lysosomes and peroxisomes, all of which float in an intracellular fluid. (5)
Cells, like human beings and other organisms, are distinguished from wooden boxes, blocks of ice and similar chunks of dead matter by the fact that they are unified causal systems which are relatively isolated from their surroundings.
In order for an entity to constitute a relatively isolated causal system in the sense here at issue, it must be the case that:
7. The external boundary of the entity is established via a covering or membrane which extends continuously across all or almost all of its surface ('almost all' because it typically contains small apertures - such as the mouth or nostrils - which allow interchange of substances such as air and food between interior and exterior).
8. The events transpiring within the interior of the entity are subject to a division between those falling within a certain spectrum of allowed values, and those falling outside this spectrum. The latter are distinguished by the fact that they will, in cumulation, lead to the entity's ceasing to exist. Almost all of the events within the interior of allowed values fall within a restricted family of types of sequences of events (for example digestion), which are often cyclically repeated.
9. The external membrane or covering serves as a shield to protect the entity from those causal influences deriving from its exterior which are likely to give rise to events which are outside its spectrum of allowed values.
10. The entity contains within itself mechanisms or procedures which are able to maintain (or, in cases of disturbance, to reestablish) sequences of events falling within the spectrum of allowed values. The entity also contains within itself mechanisms or procedures for protecting and for reconstituting its external membrane or covering in case of damage.
To see the difference between the two concepts of substance and relatively isolated causal system we note that an amoeba, towards the end of a process of fission, is one substance (by the connected exterior boundary criterion (4.)) but two causally isolated systems. It becomes two causally isolated systems gradually, via a continuous series of changes. It becomes two substances instantaneously, via an abrupt topological change. Thus the concept of a relatively isolated causal system is not a categorial concept but rather a concept that is instantiated to different degrees by different entities. A block of ice is a substance, but it is not a causally isolated system. An orbiting space-ship is a substance and, with its sophisticated mechanisms for self-repair, a causally isolated system. A circus is a causally isolated system (from the beginning to the end of a performance) to the degree that it is shielded by a membrane (the circus tent) from exterior influences. But it is not a substance, because it fails to satisfy the condition of possessing its own external boundary which separates it from other substances: the ground beneath the circus is a boundary which it must share with the earth beneath.
3. The Hierarchical Structure of the Human Organism
The concept of a relatively isolated causal system has been treated of extensively by the Polish ontologist Roman Ingarden, who points out that organisms, in order to be able to sustain themselves effectively as identical through time, must be at least in some respects 'bounded off from the surrounding world and partially isolated or, better, shielded from it.' (6)
As Ingarden also points out, each multi-cellular organism is a relatively isolated causal system which is organized in modular fashion in such a way as to contain within itself numerous further relatively isolated causal systems on successively lower levels. They are hierarchically ordered and at the same time both partially interconnected (they collaborate in their functioning) and also partially segregated from each other via coverings or membranes which protect these interiors from certain external influences and also allow other kinds of influences and substances to pass through them. The whole body, too, 'is surrounded by a well-defined enclosure - for man this is the skin, for the majority of animals - the hide.' (7) The skin or hide is itself a complex organ which is composed of many layers and has many functions in the life-process, precisely because it forms the boundary between the body and the external world. Thus it is a permeable membrane, which participates in the expulsion of water and waste-products.
Another example of a relatively isolated causal system within the body is the digestive system which, thanks to its walls, isolates ingested nutrients from the remaining parts of the body in such a way as to allow the initial digestive processes to proceed. Its permeable membranes then make possible a selective migration of the chemical products of these processes through the walls and into the blood. The heart and lungs, too, are separated from each other by appropriately constructed membranes (pericardium, pleura), which shield the processes occurring within them from outside influences.
These systems are not absolutely closed off from each other. Rather, again, they are partially open and partially shielded. There are paths between them along which a certain restricted spectrum of causal influences and substances may flow. Compare the way in which each sense organ is a partially open system which is 'attuned to a special selection of outside processes and at the same time also shielded in other respects.' (8) (1984, p. 94)
4. When Does the Human Being Begin to Exist?
We shall assume, for the moment, that everything which satisfies conditions 1.-10. above and is of human descent is a human being. (We shall find reason to add two further conditions in the sequel, but these are at this stage not pertinent to our argument.) The apparent circularity in this statement - marked by the occurrence of the term 'human' in the phrase 'is of human descent' - is innocuous given that for many purposes we can appeal to the presence of DNA of a certain specific type and this is something which can be established via quantitative means quite independent of the sorts of ontological considerations advanced here. There may, as we shall see, be other entities - such as Siamese twins - which are also human beings but which fail to satisfy one or other of the conditions mentioned. Consideration of such cases will throw useful light on the concepts we employ, but our job will be done if we are able, on the basis of the given set of conditions, to determine when these conditions are first satisfied by a developing foster in the normal course of fetal development.
We take it as unproblematic that there is a chain of events, starting with the movement of the sperm towards the egg and leading on, in normal cases, through birth, and on still further through the acquisition of self-awareness. We also take it as unproblematic that this chain of events in its early phases does not yet involve a human being and in its later phases does involve a human being which then preserves its identity through time. This proposition is so strongly supported by our ordinary pre-philosophical intuitions that one would need extraordinary arguments in order to refute it.
We shall seek to demonstrate that there is a process of substance formation of which the human being is the result, and that the formation of the human substance is a real change, occurring at a determinate time. Thus we shall reject views according to which the presence or absence of a human being might be a matter of mere convention or fiat (for example views according to which the term 'human being' expresses something like a legal fiction). We shall reject also that family of views according to which human beings are formed via the acquisition of some extra feature (such as consciousness, reason, self-awareness, the ability to make morally relevant choices) on the part of some entity which, before the acquisition of this feature, is not a human being. In case the the extra feature is acquired in virtue of some underlying causally relevant change in the entity itself, then it will be sufficient to investigate this change in the foster from the perspective of substance formation, and the same will apply also in case the extra feature is acquired via some external agency (for example God implanting a soul in the newly formed zygote) in such a way that this brings about a causally relevant change in the entity itself.
Hypotheses according to which the human being comes into existence via some external agency whose intervention brings about no such change, we can here leave out of account, since we are convinced that, armed with sufficiently powerful ontological tools, it is possible to resolve the question here at issue already on the basis of an examination of the biological (causally relevant) aspects of the matters in question.
If at some earlier time in the course of the development of the foster a human being does not exist, and at some later time a human being does exist, then at some point in the intervening interval a change takes place which is a categorial change. We said earlier that a substance is a bearer of change. We now see that substances cannot be the sole bearers of change, since a categorial change is a change in which a substance itself comes into or goes out of existence. Consider a larva which turns into a butterfly. The same matter now instantiates one form, now another. This suggests that in order to do justice to the fact that change occurs only if there is something which changes we adopt the view developed by Aristotle in his Metaphysics. According to this view, a categorial change is a change in matter, and the identity through time and survival of a substance, for example of an organism, is based on the continued existence of a certain form or organization. (9)
Consider two pools of water which merge to become one. The same matter now instantiates one form, now another. Something similar must occur in the course of development of a human being. A given body of matter now instantiates one form (say: that of a cluster of cells), now another (it is a human being, a single substance). Such a change of category is different from the ordinary sort of change which you undergo when you dye your hair or grow your nails. With a change of category one entity ceases to exist and some different entity comes into existence.
5. The Varieties of Substance Formation
We know from other domains that the formation of entities can happen in a variety of different ways. Biological species form via budding from existing species. Bacteria form via cell division. Bangla Desh was formed via the separation of East and West Pakistan. Budapest was formed via the fusion of Buda and Pest. How do human beings form? Let us outline here the possible alternatives:
i) Budding: A part of one individual substance is detached therefrom and forms a new individual substance in its own right while the original substance goes on existing. We might detach the tail from a cat. Before the loss of the tail, the cat is both one substance and one relatively isolated causal system. Upon the loss of the tail, which is an abrupt topological change, both the cat and the tail are substances. However, only the cat is a relatively isolated causal system; the tail is just dead matter. Substance formation of a biologically more pertinent sort occurs through budding in those organisms which reproduce by making small dormant cells within themselves and then releasing them into the environment.
ii) Separation: Two or more entities are joined together as one entity and at some point the relations conjoining the parts of this entity are disrupted in such a way that the previously attached individuals continue as separate new substances. Consider, for example, a pair of Siamese twins who are separated by means of surgery. The connected Siamese twins constitute one substance, by our criteria, where the separated Siamese twins are two substances. The Siamese twins, however, may constitute two relatively isolated causal systems both before and after the detachment.
iii) Fission: Fission is distinct from separation in that, when an entity (for example a virus or other microbe) undergoes fission, new parts are formed which then split apart to lead separate existences. Fission gives rise to new entities and destroys the entities which existed earlier. An amoeba, for example, divides by duplicating its nucleus and then allowing the rest of the cell to split apart, via an abrupt topological change, into two new organisms of roughly equal size.
iv) Unification: Separate individual substances join into a complex which forms one new substance in which the previously detached separate substances continue to exist as entities joined together within the new whole (unification is the dual of separation). Many examples of this type of substance formation are found in the realm of physical processes (for example when separate water molecules condense to form a single drop of water) or in the realm of artefacts (when a table is formed by nailing together several blocks of wood).
v) Fusion: Fusion is distinct from unification in that, when entities undergo fusion, then they thereby cease to exist (fusion is the dual of fission). Fusion gives rise to a new entity, as for example when two drops of water join together to become one; their respective parts merge to form a perfect mixture.
6. The Development of the Foster
Our aim is to establish the ontology of the process by which human beings are formed. To this end we will need to consider briefly the biological details of the development of the foster.
The story begins when an egg-cell, developed in the ovaries, is released into the fimbrated end of the fallopian tube (ovulation). The egg-cell, swimming free in the fluid-filled tube, is encountered by a sperm, and the latter initiates a process of penetration. (This is what happens in the normal case. In very rare circumstances a sperm cell might attach itself to one of the additional, much tinier cells which are also floating in the fallopian tube. These are the so-called 'polar bodies', deriving from eggs released at earlier stages in their development.)
Fertilization. The first stage in the process of penetration is the fusion of the membrane of the sperm cell with that of the egg. This results in the formation of a channel that allows the passage of the nucleus of the sperm cell into the egg. The male genetic material that is carried by this nucleus then fuses with genetic material from the nucleus of the egg-cell. The two nuclei come into contact in the egg cytoplasm, and shed their nuclear membranes. Each offers up one complete set of 23 chromosomes, and these two sets of chromosomes become entwined around each other as part of a process which transforms the egg-cell into a new joint product, called the zygote, an unusually large cell which has the same membrane as the egg-cell before fertilization.
Cell Division. Immediately upon formation the zygote begins to undergo a process of cell division. During the very first cell divisions (up to the eight-cell stage), there is no qualitative distinction between the cell that is dividing and the cells resulting from the division. The cells are functionally and qualitatively undifferentiated. Thus, each has the potential to produce a complete human being (each is, as the jargon has it, 'totipotential'). The cells are kept together spatially by a surrounding thin membrane (the zona pellucida), which was also the membrane of the egg-cell before fertilization, but there is no causal interaction between them. They are separate bodies which adhere to each other through their sticky surfaces. The cells are at this stage still floating free, inside their common membrane, in the fluid-filled fallopian tube, but they have begun to move towards the womb (uterus).
Formation of the Morula. Because of the limited space within the zona pellucida, a compaction takes place between the 8- and 16-cell stage (day 3). (10) As a result of this compaction, the inner cells divide faster than the outer cells which surround them. At this time tight junctions between the cells are formed; neighboring cells are connected by highly selective permeable membranes which are marked by channels through which signal molecules can be transported from one side to the other. The resulting bundle of cells is now called a morula (Latin for 'mulberry'). The morula is formed as the cells move from the fallopian tube and into the womb.
When there are about 60 cells present (day 4), there occurs a clear differentiation between the so-called 'inner cell mass' and 'the trophectoderm' (an outer ring of cells). The latter functions henceforth as the surrounding membrane in place of the zona pellucida which disintegrates. In addition, pools of clear fluid which had accumulated between some of the internal cells coalesce to form a common cavity called the blastocoel, a body of extracellular fluid in which the inner cell mass thereafter floats. The two groups of cells, taken together with this fluid, are now called the 'blastocyst'. The entire blastocyst itself floats freely in the uterine fluid for about a day and continues to exist disconnected from the mother.
Implantation. Over the next week (days 6-13) there occurs a process called implantation (also 'nidation', from the Latin nidus: a nest, or niche). The blastocyst, on completing its journey along the fallopian tube into the uterine cavity, moves into a position where it is in contact with the uterine wall, to which it adheres via its sticky exterior. Cells on its outer surface then begin to grow rapidly in such a way as to disrupt the surface of the wall. These cells actively burrow into the deeper tissue until they have become completely embedded. The inner cells of the blastocyst are however still not connected to the mother, since they float in the liquid contained within the trophoblastic membrane. With the implantation of the blastocyst in the wall of the womb comes the formation from its inner cell mass of what is called the 'embryonic disc'. This consists of two kinds of cell mass: the epiblast, which will eventually give rise to the embryo proper and to parts of the umbilical cord; and the hypoblast which will give rise to extra-embryonic membranes and tissues.
Gastrulation. When the embryonic disc is fully implanted in the wall of the womb (day 13), it is able to grow and can begin to differentiate into the various tissues that will characterize the human being into which it develops. Only at this point does the disc begin to receive nutrients from the mother. Until now, only cell division has taken place. The morula has not grown in size compared to the egg-cell; rather, its constituent cells have become smaller. Now however the embryonic disc begins to grow, as the neural folds are formed and the primitive cardiovascular system begins to function. This stage, which begins around day 14-15, is called 'gastrulation'. It consists in a convergence of the epiblastic cells into what is called 'the primitive streak', a stacking of cells on the end of the embryonic disc which will develop into the spine. During the process of gastrulation (days 14-19), the epiblast differentiates into three cell types: the mesoderm (which will become the smooth muscle coats, connective tissues, blood cells, bone marrow, the skeleton and the reproductive and excretory organs of the fetus), the ectoderm (which will become the epithelia and the nervous system), and the endoderm (which will become the linings of the respiratory passages and of the digestive tract). In addition, some cells of the epiblast form the umbilical cord. Not all of the cells of the embryonic disc are going to develop into the fetus. Some of its cells (the non-epiblastic cells) will form the extraembryonic membranes (the amnion, the chorion and part of the placenta) and other extra embryonic tissues.
The embryonic disc is now commonly referred to as the embryo proper, a term which is used to describe the developing foster until the ninth week after fertilization, from when it is called the 'fetus'.
A few days after gastrulation there begins to form the fluid-filled amniotic cavity in which the foster will float until the end of its term. This amniotic cavity is within the wall of the uterus, and as it expands it brings about a consequent contraction of the uterine cavity proper. Surprisingly, therefore, the embryo/fetus is for almost the whole of its development not, strictly speaking, inside the uterus (or womb, or uterine cavity); rather it is lodged within a specially created cavity inside the uterine lining. This multiple-cavity structure enclosing the foster provides a cushion against mechanical injury.
The placenta is a flat organ which supplies nutrients for the fetus during its development. It develops from the outer cell layer (from the trophoblastic cells) in the early embryo which fastens itself to the wall of the uterus from around day 21 and becomes anchored to the mother via a maternal portion formed by part of the functional layer of the uterine membrane. Together with the other fetal membranes (the amnion and the chorion), it disintegrates some moments after birth, when it is delivered through the birth canal.
The umbilical cord is an organ of the fetus that penetrates the placenta via two large arteries which radiate outwards from the point where breaks through into the inner surface of the placenta and divide into small arteries that penetrate ever further into the depths of the placenta through hundreds of branching strands of tissue known as 'villi'. These villi cause a rupturing of the mother's blood vessels in their vicinity and are thereby bathed in maternal blood. The constant circulation of fetal and maternal blood and the very thin tissue separating fetal blood from maternal blood bathing the villi provide a mechanism for interchange of blood constituents between the maternal and fetal bloodstreams. However, it is normally not the case that there is opportunity for the blood of one to gain access to the blood vessels of the other. Rather, nutrients, oxygen, and antibodies diffuse into the fetal blood in the capillaries of the villi, and wastes and carbon dioxide diffuse out of these capillaries into the maternal blood circulation. (Compare the way in which oxygen is transmitted to fish via pipes which feed air into the water of an aquarium.)
Development of the Fetus. At about 40 to 43 days after conception the rudimentary brain at the top of the primitive streak begins to function. At the ninth week, the fetus has almost all human characteristics (except for the face and genitals) and it begins to show signs of specific male or female development. During the tenth week, the face and the genitals begin to develop. In the twelfth week, when the foster is nine centimeters long, it begins to move its hands and feet. Its movements increase around the sixteenth week, when hair also begins to grow and teeth are developed. At the twentieth week, the foster can suck and swallow and its body bends and stretches. From then on the foster continues to grow in size until, at the end of the fortieth week, it is born.
7. Alternative Thresholds
Given this simplified account of fetal development, let us return to the question of transtemporal identity. When does the foster first satisfy the ten conditions set out above for being a substance which is also a relatively isolated causal system? At what stage is the foster transtemporally identical to the human being it becomes after birth? We distinguish the following nine possibilities:
a. The stage of the single-cell zygote (day 0)
b: The stage of the multi-cell zygote (days 0-3)
c. The stage of the morula (day 3)
d. The stage of the early blastocyst (day 4)
e. Implantation (days 6-13)
f. The formation of the primitive streak (days 14-16)
g. Viability (e.g. day 140)
h. Birth (day 275)
i. The development of self-consciousness (some time after birth)
a. The zygote is a substance: it is a bearer of change; it persists through a time-interval; it is extended in space and it has spatial parts such as the nucleus, the cell-membrane and the filaments inside it; it has its own connected exterior boundary which divides its interior from its exterior and which connects the parts within its interior and thus distinguishes it from a mere heap or collection. Moreover the zygote is an independent entity in the sense that it does not require the existence of any specific second entity in order to exist. (Thus it can survive transplantation.) The zygote is moreover, like every other cell, a relatively isolated causal system. It is shielded by its outer membrane from causal influences deriving from its exterior; the events transpiring within its interior are subject to a division between stable and critical events; and it contains its own rudimentary mechanisms for reestablishing stability in cases of disturbance. But there is an obvious reason why this zygote substance cannot be identical to the human being which will exist after birth, namely that it is predestined to undergo fission, and this means that it will cease to exist almost immediately after it has been formed.
b. The second possibility is that it is at the stage of the multi-cellular zygote-bundle that the human being comes into existence. But the zygote is at this stage most properly conceived as a sticky assemblage of 8 or 16 entities rather than as a single entity. They are not one but many. Although they are surrounded by a thin permeable membrane, this membrane merely helps to keep the cells together in the spatial sense; there is no flow of nutrients or signal molecules from the outside to the inside of the membrane or from one cell to another, and the cell bundle has no stability-restoring mechanism of its own of the sort which is required in order for the whole entity to be a single causal system.
Perhaps, though, we can hold on to the view that the multi-cell zygote is already a human individual by arguing that some one cell within the bundle is privileged by the fact that it inherits from the original single cell the property of serving as the bearer of identity for the human being that is in process of development. The problem with this view is that it contradicts totipotentiality -- the feature in virtue of which each of the cells within the multi-cell zygote has the full potential to develop into a human being.
To see the problem here, we must understand how differentiation works. Differentiation is the creation, from a mere aggregate of homogeneous cells, of clusters of functionally and structurally different types of connected tissue at different sites. In the case before us, all the cells maintain forever the same genetic composition (that of the original fertilized egg-cell). However, the very genes involved contain the programming for differentiation. This programming goes into effect in different cells in different ways, not because of any intrinsic features of the cells themselves, but rather as a result of the specific environments surrounding them and thus of the macroscopic structures which they together go to form. This surrounding context determines that some of the genes within each given cell become repressed, so that only some types of protein are made. That it is the environment surrounding a given cell which determines what kind of proteins will be formed (or 'expressed') by the cell can be seen from the fact that, if cells of a given type are moved artificially to a different environment where they are surrounded by cells of a different type, then they will begin to express the same proteins as the cells which surround them. Since, at the stage of the multi-celled zygote, no differentiation has taken place, it follows that there can be no cell or bundle of cells within the cluster which is privileged in virtue of some intrinsic feature which it might possess.
c. At the formation of the morula, too, differentiation has not yet taken place, and so the just-mentioned argument can be applied in this case also (as also in cases d. and e., below). Each of the cells of the morula still has the potential to become a human being. At this stage, junctions between the cells of the zygote are formed which allow intercellular communication by means of small signal-molecules. But the morula still fails to meet condition 10. for being a causally isolated system. That is, it does not possess mechanisms of its own to restore stability in cases of external disturbance. At best it must rely on the separate rudimentary stability-restoring mechanisms of its constituent cells. The morula is still, in our terms, a loosely connected cluster of cells.
d. At the stage of the early blastocyst, the cells have differentiated into the inner cell mass and the surrounding trophoblast. The inner cell mass constitutes a single substance, rather than many substances, insofar as its cells together form a single connected whole with a common physical boundary. The inner cell mass also begins to constitute a relatively isolated causal system, but it still lacks its own internal mechanisms in virtue of which its several parts would in cases of disturbance work together as a whole to restore stability. The inner cell mass will differentiate into two further tissues and only one of them will eventually become the embryo. The other will turn into membranes and extraembryonic tissue. This is not in and of itself important for determining whether or not it is transtemporally identical to the later human being, for one may argue that the membranes and extraembryonic tissues are merely temporary parts of the embryo in much the same sense as baby teeth are temporary parts of the child. What is important, however, is that, following our account of differentiation above, it is not yet determined which parts of the inner cell mass are predestined to become embryonic cells. Thus the stage of formation of the inner cell mass at day 4 still does not seem to be a good candidate stage for the formation of the human being.
e. When the process of implantation comes to an end, the embryo can begin to receive nutrients from the mother and can begin to grow as an individual and to differentiate into tissues of different sorts which are recognizable precursors of neonatal tissues. However, as for the early blastocyst, so also here, it seems that the foster still lacks its own integrated mechanism for restoring stability, and so it fails to be a relatively isolated causal system in the sense at issue. An identification of the foster as it now exists with the later human being, however, faces the problem that the foster has entered into a condition of being dependent on the mother the mother for nutrients and oxygen as well as for temperature control and other forms of protection. Does this imply that the foster is henceforth such as to fall short of being a substance because it does not satisfy condition 6.? Certainly, if it is extracted from the mother it will almost certainly die through lack of nutrients and through lack of an appropriate protective environment into which it can fit. But this applies also, for example, to an Arctic explorer in relation to her winter ice station. It is known from the animal kingdom that premature offspring can often survive in external environments. A kangaroo foster, for example, is born alive at a very immature stage when it is only about one inch long and weighs less than one gram. After birth it uses its forelimbs to crawl up the mother's body and enter the pouch which is a pocket on the mother which opens forward and contains teats. The baby kangaroo grows in the pocket and gradually spends more and more time outside the pouch, which it leaves for good at the age of seven to ten months. The amniotic cavity, in which the foster develops upon implantation, is in some sense like a kangaroo pouch, though instead of being open (in the sense of having an aperture) it is a closed cavity. As the kangaroo foster is not rigidly dependent on one specific mother, but only on an appropriate environment with teats and so forth, so the foster is not rigidly dependent on the mother, but only on some appropriate environment, which might from a certain point be supplied by means of an incubator. This means that the foster is not dependent on the mother in the ontological sense of dependence that is involved, for example, in the relation between a smile and a face, or between an individual instance of color and some extended surface.
f. According to Ford (1988), the point where a human individual comes into existence is the time when the epiblast ceases to be a cluster of homogenous cells and begins to transform itself into a heterogeneous entity. He suggests that this happens with the formation of the primitive streak (around day 16). At this point, he argues, the cells of the epiblast have formed a whole multicellular individual living being which has a body axis and bilateral symmetry; that is, it is spatially oriented. The embryo's cranial axis and its dorsal and ventral surfaces come into existence, and the whole has acquired effective mechanisms to protect itself and to restore stability in face of disturbance.
McLaren (1984), too, offers a defense of the thesis that it is the formation of the primitive streak which marks the beginning of the human being. This, he argues, marks the beginning of a human individual insofar as the boundaries of a discrete, coherent entity have now been formed. At this point the foster also meets the other part of criterion 8. to the effect that it is protected against outside disturbance by the jacket of cells surrounding it. That is to say, at the formation of the primitive streak, there is formed for the first time a bona fide spatial boundary which delineates the embryo spatially from the extraembryonic tissue. Before this point it is not determined which cells will become the embryo and which the surrounding membranes. In the case of adult human beings, in contrast, it is determined which parts (for example: nails, hair) can be removed without the human being ceasing to exist. In this respect too, therefore, the foster is from the stage of the formation of the primitive streak more like the adult human being than it is like the foster at earlier stages.
The tighter integration of the foster that arises with the formation of the primitive streak is manifested most importantly in the fact that twinning is from this point on no longer possible. If fission occurs just before the formation of the primitive streak, this will in almost all cases give rise to malformations (Siamese twins).
g. It has often been suggested that the human individual begins to exist at the point when the foster becomes viable; that is, when it can live outside its mother's womb. The argument is that because, prior to this time, the foster cannot survive independently of its mother, it is analogous to an organ of the mother, which can only exist and exercise its proper function within the locus of its proper encompassing environment. The problem with this view, however, is that the transition to viability does not in itself connote a transformation of one entity into another. The passage into viability of the foster may coincide with such a transformation process, but it does not necessarily do so. Rather, this passage may involve no genuine change in the foster, but may represent a mere Cambridge change of the sort which is expressed by propositions such as 'John is no longer beloved in South Africa', or 'Mary just ceased to be the tallest player in the team.' Acquiring stronger muscles is a real (though not a categorial) change, but the satisfaction of the viability criterion is not dependent on such physical changes in the foster; it may be satisfied, for example, through advances in technology.
h. Another alternative is that it is the event of birth which marks the initiation of the human being. Consider, for example, the Talmudic doctrine according to which the foster is a limb of the mother, so that only once its head has emerged from the mother's body does it begin to exist as a substance in its own right. We shall argue below, however, that birth is the mere passage of an entity from one environment to another. That is to say, it is a process of a sort that does not affect the ontological category of the entity involved. If the human being exists at birth, then it exists also in the minutes prior to birth, and then our question as to when the human individual begins to exist arises once again.
Could birth constitute a sharp boundary for the initiation of human existence? Bermudez (1996) points to the fact that there is a difference between the fetus and the newly born infant in reflection of the fact that the neonate can engage in forms of cognition that are unavailable inside the womb. Consciousness emerges outside the womb, and is largely dependent on the extra-uterine environment. On the view we have been defending, birth does not involve a transformation of the foster into some new substance. But might we not take an alternative view, according to which the concept human being is not a substantial concept at all, but rather a relational concept, so that to be a human being is to be, inter alia, in a certain sort of environment? (Compare the way in which the concept of a pawn is tied to the environment of a chess game.)
The objection to any approach along these lines is that it is not clear whether the thesis that a human being exists only in a certain sort of environment is a conceptual thesis or a merely empirical thesis. If it is conceptual in nature, then we can consider a thought experiment in which a fetus remains in a specially tended and adapted womb/incubator perhaps for many years beyond its term, growing and acquiring many of the features normally considered as characteristic of a human being, perhaps including features of reason and consciousness. The organism in question would not be a human being, on the relational understanding. If, on the other hand, it is an empirical thesis, then it is unclear what is to count as evidence for or against its truth.
i. The final alternative is that it is the acquisition of some extra feature which marks the beginning of the human individual. One obvious candidate is consciousness, and in particular self-consciousness. This position was, for example, supported by Locke, for whom the main characteristic of a human individual is that it can consider itself as itself. (11)
Can we, then, identify the acquisition of the capacity for self-consciousness (or of some similar capacity) as marking the point when the human being begins to exist? Such a dispositional property requires in every case some underlying real basis, i.e. a basis which is not itself merely dispositional. The corresponding physical change in this case might be an increase in complexity of nerve-connections in the brain. The proponents of alternative i., therefore, are asking us to accept that a change of this sort within a certain part of the matter of the object would bring about a substantial change in the object as a whole. We do not rule out the possibility that a thesis along the lines might be true. However if it is more than a mere conceptual or definitional matter, then the thesis would require powerful argumentsof a sort which have not been provided. In other cases where a new capacity is acquired we do not, after all, talk in terms of the corresponding underlying physical change in a part of the organism as marking the coming into being of a whole new substance. (12)
8. Twinning
The principal argument against alternatives a. - e., above, turns on the phenomenon of twinning. Ford (1988) and Diamond (1975), among others, have argued that the zygote/early embryo cannot be a human individual if it can undergo division in such a way that the results of this division can develop separately to form two distinct human individuals. A single human individual cannot give rise to two human individuals in this way. One argument in support of their view might read as follows:
A correct account of the concept human being must in every case yield a determinate answer to questions like 'how many human beings are there in this room?' The possibility of twinning shows, however, that no concept satisfying this condition could be applied to a foster before the emergence of the primitive streak, because prior to this stage of development it will be indeterminate whether there is one or two individuals.
To see the force of this argument we need to focus carefully on the question of the transtemporal identity of the foster with the human being which exists after birth in the case where twinning does not in fact occur but is nonetheless still possible, in light of the three alternative accounts of twinning which can be distinguished in light of the three varieties of substance formation by division - budding, separation and fission - distinguished above.
A proposal defended by Daly (1985, p. 51) is that the human individual can survive the twinning process because the latter can be understood as a process of forming human individuals via budding. Twinning, on this account, is in fact a form of cloning. Similar phenomena are after all known from the vegetable kingdom, where a cutting from one plant may be planted in the soil to result in a new plant without the original plant ceasing to exist as a separate individual. Daly also mentions mitosis in bacteria, where a single bacterium divides into two separate bacteria. However, neither of these examples is convincing as examples in favor of the thesis that twinning among humans is consistent with the continued existence of an individual substance. First, human development is nothing like that of plants. When a cutting is taken from a fully developed plant we do not have one cell which divides into two; no division takes place at all. Rather the situation is analogous to one in which one would grow a new human individual from a nail or a lump of hair. The bacterial example is more convincing in this respect, but the problem here is that nothing in bacterial mitosis supports the view that the original bacterium survives as an individual substance. Rather, everything points to an account of the matter according to which bacterial division is a process of fission in our technical sense.
There is another problem with the view proposed. Consider the so-called 'mosaic chimeras' or twin-within-a-twin individuals. (13) These are single human individuals, most of whose tissues bear the marks of two distinct cell lines which come from two distinct sperm (they may even, in rare cases, come from two distinct fathers). Mosaic chimeras result when one sperm fertilizes the egg and another fertilizes one of the other cells (the already mentioned 'polar bodies') separating at the time of the formation of the egg. The two zygotes may then fuse to form a single individual (which is accordingly a result of quadruple fusion). Suppose we were to say, with Daly, that zygotes are already human individuals. We would then have to say, in the case where both the egg-cell and a polar body are fertilized, that the two fertilized cell clusters are already two human individuals. A mosaic chimera is the result of fusion of two such fertilized cell clusters. In all other respects, however, it is like every other human being: it is one substance and one relatively isolated causal system, and it has all the characteristics of a normal human individual.
What of the case where a foster is structured in such a way that twinning can still occur via separation? This would mean that the foster is already not one but two entities, both of which would survive, should twinning occur, to form two independent fosters. The separable foster, at the times when separation is still possible, would lack the sort of integrity which would make it a unity in its own right. It might be compared to a pair of railway carriages joined together by a loose bond, which may break or which may be strengthened by degrees in such a way that the results of this strengthening form a single entity. The problem with this view, however, is that we know empirically that twinning is, at least in many cases, a phenomenon that is caused by accidental interventions deriving from the external environment; thus it is not in every case determined at times prior to division that twinning will occur. Thus an account of twinning as separation provides nothing to which we could point as the human being at stages when twinning of this accidental sort is still possible.
What, finally, of the case where twinning can still occur via fission? Here we might suppose that we have a structure analoguous to that of an amoeba, whose interior bonds are insufficiently strong to prevent division, but which are yet sufficiently strong to constitute a unity. In such cases, surely, we would have no hesitation in asserting that the identity of the entity in question persists across an interval during which the amoeba is susceptible to fission but does not, in fact, divide. The case is however much rather one in which the interior bonds of the foster undergo a cumulative process of strengthening. Until this process has reached a point where twinning is no longer possible, we have a situation that is analoguous to a collection of railway carriages that might divide and coalesce into separate clusters of carriages in any one of a large number of ways, each one of which would signify the ceasing to exist of the entity that is thereby subject to division. Here again, therefore, we have nothing to which we could point as the human being at stages when twinning is still possible.
It follows from the above that we need to add a further condition to our list above in order to rule out from candidacy as human beings those entities which have within themselves the potentiality to undergo division. All organic entities lose parts over time (as you lose hair and skin, for example). What we need to exclude is that an entity of a given type should be subject to a division into two or more entities which are similar to itself in the way in which this is possible, for example, in the case of an amoeba, or an earthworm. An entity is non-divisible, in this sense, if it satisfies:
11. The parts of the entity and the processes in its interior are integrated together in such a way that it is not possible that the entity should divide into two or more entities which themselves satisfy conditions 1.-10. above.
An entity which satisfies condition 11. is, we might say, a unity, a single substance, not merely actually but also potentially.
9. The Concept of Niche
But we must add one further condition to our list. For consider John's heart. This is a substance, and it is a relatively isolated causal system; it is non-divisible; and it is a product of human reproduction; yet it is not itself a human being because it is not a maximal entity satisfying these conditions: it belongs as proper part to John's organism as a whole. On the other hand, however, John is himself not a maximal causally isolated substance in all the phases of his existence. Suppose, for example, that John is inside a spaceship and is working the ship's controls. The mereological sum of John and the spaceship is then a substance, by our criteria above. It is also, to a degree, a relatively isolated causal system. But there is however a difference between John in his relation to the spaceship, on the one hand, and the various cells and organs in their relation to the organism of which they form a part, on the other. John is not in the spaceship in the way in which, for example, a nucleus is in the cell or the heart or brain is in the body. Rather, John is in the spaceship as a bird is in its nest or as a bear is in its cave. More generally, John is in his spaceship as an organism is in its niche (which means, inter alia, that he can leave the spaceship and then return, he can be replaced by another human being at the ship's control, and so on).
Intuitively, a niche is a part of reality into which an object fits, and into and out of which the object can move. A niche and its tenant do not overlap (they have no parts in common). Rather the niche surrounds its tenant, and the notion of surrounding that is here at work can be precisely specified via tools of mereology, topology and the theory of spatial location. (14) A niche and its tenant are disconnected from each other in the sense that they share no boundaries in the way in which, for example, a cat(-torso) shares a boundary in common with its tail. The surrounding relation between a niche and its tenant is thus to be distinguished from the relationship of perfectly tight connection of the sort which we find in the case of David trapped within the interior of the block of stone before Michelangelo gets to work. For there must on all sides be a certain degree of free play between the niche and the tenant housed inside it. The niche-tenant relation must therefore involve some sort of cavity, of air, water, or some other medium, by which the tenant is surrounded. John, in sum, is in the spaceship not as a part is in its whole, but rather as an organism is in its niche.
To see the contrast here, let us suppose that John is involved in an accident which destroys the skin over the entire surface of his body. On one scenario, the doctors graft a new skin onto his body that is made of some synthetic material. On another scenario they create a space suit-like cover for John, which he will henceforth wear. In the former case, we say that the synthetic skin is a part of John (the doctors have created for John a new, synthetic organ). In the latter case, we say that the space suit creates a niche for John, into which John fits, and which is then such that John and his niche have no parts in common. Note that the issue here does not turn on what the skin, or suit, is made of. We can imagine, for example, some future spacesuit-like container-niche for John that is constructed out of human DNA teased into weavable plastic form.
We can now lay down our needed supplementary criterion as follows. Intuitively, we want it to be the case that a maximal entity in the sense here at issue is an entity which, if it belongs as part to a larger substantial whole, then only because it stands, within this larger whole, in a tenant-niche relationship. We can state, more precisely, that an entity x is a maximal substantial, non-divisible relatively isolated causal system if:
12. (i) x satisfies conditions 1.-11. and (ii) every entity y which has x as a proper part and which also satisfies these conditions has some part which is a niche for x.
Normally, the niche for x will simply be the result of subtracting x from the whole y. Thus if John (x) is alone inside the spaceship, then the relevant value of y is: John plus the spaceship (the two mereologically summed together), and the relevant niche is just the spaceship itself. John is not a part of this niche, but rather (trivially) a part of the mereological sum of the niche together with himself.
An interesting illustration of the workings of our twelve conditions is provided by the case of cryogenically frozen human beings. We are to imagine a living human being whose metabolism is suppressed by freezing and who is preserved in the frozen state by means of some surrounding refrigerator-like mechanism. Here, the organism's own mechanisms for sustaining and restoring bodily stability are put out of action and the tenant of the cryogenic niche borrows these mechanisms from its new artificial surroundings. Yet the organism still has these mechanisms (even if in a dormant state), and thus our twelve conditions are fully satisfied. It is an implication of the view here defended that brain-stem death does not in and of itself amount to the ceasing to exist of the human being. The brain-stem dead individual is, precisely, kept alive in an oxygen cubator and satisfies all the conditions set forth above. However, our twelve conditions may not be satisfied by human beings on the threshold of death, beings who are dying precisely because the mechanisms for self-repair are no longer functioning. To some degree in conflict with current medical orthodoxy we prefer to describe such an individual as being not dead, but rather (irreversibly) dying.
Our remarks on cryogenically frozen human beings can throw some light also on the question whether human beings in general might be formed at some very early stage in their development (e.g. at the stage of the single zygote) though in such a way that the capacities which make them relatively isolated causal systems would exist only in a dormant form. Some marginal kinds of life (e.g. a virus in its non-active phase) may display for long periods a complete lack of metabolism, but we call a virus living because of its potential to engage in an active, self-replicating phase when it is coupled to the metabolism of more complicated host organisms. The same goes for simple animals (e.g. some shrimp species) which are able to stay completely metabolism-free during long, cold seasons. Such organisms must have the capacity to display metabolism in other phases of their existence, otherwise we would not call them organisms at all. What counts against drawing conclusions from such considerations for the human case, however, is once again the issue of twinning, which is still possible at the stage of the zygote and at all later stages prior to the formation of the primitive streak.
10. Is the Foster Connected to the Mother?
Any entity which satisfies condition 12. and which is a product of human reproduction will be a human being. From the stage of formation of the primitive streak (around day 16), the foster constitutes a single substance with its own external physical boundary, it is a relatively isolated causal system with its own mechanisms for restoring stability, and it is non-divisible. It remains to show that the foster from this stage satisfies also the maximality condition. We need to establish, in other words, that every entity containing the foster as proper part which satisfies conditions 1.-11. has, from this time, some part which is a niche for the foster. Since the only candidate for such an entity is the mother, we need to establish the nature of the relation between the foster and the mother: is it that of tenant to niche, or that of part to whole, or is it some other sort of relation?
It will help to distinguish here between 'fiat' and 'bona fide' boundaries. (15) Fiat boundaries are boundaries which do not fully correspond to physical discontinuities in the underlying territory. Examples are found above all in the realm of arbitrarily demarcated entities such as counties, postal districts, census tracts, or air traffic corridors. Bona fide boundaries, on the other hand, are physical discontinuities in the usual sense of this term: for example the edge of this table, or the surface of that painting, or the outer boundary of your heart or lungs. Recall our condition 4., above, according to which each substance must possess its own complete, connected boundary which divides its interior from its exterior. In condition 7. we asserted that this boundary is established via a covering or membrane which extends continuously across all or almost all of its extent. This means that the boundary of a substance is a bona fide boundary for all or almost all of its extent and that where the boundary falls short of bona fide status this is because fiat boundaries are formed in virtue of apertures in the shield or membrane, for example in the region of the mouth or nostrils. To say that fiat boundaries are formed is to say simply that the line between interior and exterior in the region of such apertures corresponds to no physical discontinuity. (You can create a fiat boundary of this sort, as a part of the total boundary of your office, by opening your office door.) Siamese twins, which have a part of their surface (and sometimes also certain organs) in common, are also separated by a fiat boundary along the line where they meet.
These remarks are important since categorial changes are closely connected to topological changes - to changes in boundaries and in connectedness - and such changes can be properly understood only if careful attention is paid to the fiat/bona fide opposition (Smith and Varzi, forthcoming). Two entities have the same topology, roughly speaking, if one can stretch, bend, shrink or expand the one entity, without tearing or joining, in such a way as to turn it into the second entity. Two entities will then have the same topology if and only if they have, across their surface and interior, analogous distributions of fiat and bona fide boundaries. A newly born child has in most cases (Siamese twins provide exceptions) the same topology when it is born as it does after it has turned six. A tail that is attached to the body of a cat does not have the same topology as it would have if it were detached from the cat: this is because before detachment its boundary includes a fiat as well as a bona fide portion.
Fiat boundaries are involved, now, whenever an organism undergoes a process of division. As an amoeba splits its middle part becomes progressively thinner so that we talk in terms of two halves of the amoeba with a fiat boundary running within the interior of the amoeba between them. As the middle part shrinks this fiat boundary separating the two halves then goes out of existence, incrementally, until the point is reached when it disappears entirely. A long, continuous process suddenly results in an abrupt topological change. The right and left portions split, and we have two amoebae, each with its own complete bona fide boundary.
Fiat boundaries are involved, too, whenever one entity is strictly connected to another (in the way in which, for example, the tail is strictly connected to the remainder of the cat). The tenant-niche structure, as we saw, excludes the possibility that tenant and niche may be strictly connected to each other. Thus if the foster is connected to the mother - if, in other words, the boundary between the foster and the mother is a matter of fiat and not of bona fide boundaries - then the foster cannot stand to the mother in the niche-tenant relation, and thus it cannot satisfy condition 12.
But as we have seen, there is at no stage after ovulation a strict topological connection between the foster and its mother. Such a connection is not even established in the form of a canal through which blood or nutrients might flow. (16) The communication taking place between the foster and the mother involves many separate processes of cell diffusion, but these processes occur always via some intervening liquid-filled cavity; they never involves the presence of that sort of common membrane or covering which would be required for strict topological connection. It is for this reason that birth is most properly conceived as the passage of an entity from one environment to another; it is analoguous to John moving out of the spaceship.
11. Is the Foster a Part of the Mother?
We are now in a position where we can address the question whether the foster, upon the formation of the primitive streak, is a part of the mother, or rather a tenant lodged within a niche inside the mother. Is the amniotic cavity (the cavity created by the foster within the interior of the uterine lining) itself a part of the mother? Or is it a genuine cavity, disjoint from the mother, even though lying within her interior?
If an entity is a proper part of a second substance which is a human being, then it fails thereby to satisfy our condition 12., since a tenant never shares parts in common with its niche. Suppose a tub of yoghurt is inside a refrigerator. The yoghurt is then in the interior of the refrigerator; but it is not a part of the refrigerator. Indeed it and the refrigerator share no parts in common. Rather, it is lodged within a cavity within the interior of the refrigerator, and it relates to the refrigerator as a tenant to its niche. The refrigerator is one substance, the tub of yoghurt is a second substance, and the former surrounds the latter. Topologists distinguish in this connection between an object and its convex closure (the latter is, intuitively, the result of filling all the cavities within the former). If we say: the tub of yoghurt is inside the refrigerator, then we really mean that its spatial location is a part of the spatial location of the convex closure of the refrigerator. (Casati and Varzi 1999)
Consider the analogous case of the digestive tract. Is the digestive tract a tunnel through your body, and thus a hole, or rather an organ within your body and thus a part? Certainly the throat, esophagus, stomach, and colon are parts of the body as systems of cells. But the canal which they together form is a genuine cavity, and thus not a part, since its function is, like that of the mouth, to serve as a protected antechamber which is able to withhold ingested substances from the body in order to make it possible for the initial digestive processes to proceed. And so also in the case of the complex of nested cavities in which the foster is lodged.
Certainly the foster is, from the stage at which the inner cell mass is formed, a part of that substance which is the mereological sum of this inner cell mass together with the trophoblastic cells and the blastocoel. It is also a part of that circumcluding substance which is the mereological sum of all of these together with the amniotic membrane and the fluid inside the amniotic sac (the 'bag of waters'). And it is also a part of that even more inclusive substance which is the mereological sum of all of these together with the mother. The foster is thus a part of a whole which includes the mother as part; but it is not a part of the mother. Rather, we can now conclude, it is it occupies a niche within the interior of the mother (in a way that is analogous to a body that is lodged within your digestive tract). Thus the maximality condition (12.) is satisfied. (17)
It is illuminating to consider the contrasting cases of the umbilical cord and the placenta, for these two organs seem to satisfy many if not all of the 12 conditions listed above, but they are clearly not human beings. On closer inspection, however, it becomes clear that the umbilical cord is a part of the foster in the same way in which a baby tooth is part of a child. Thus it fails to meet our maximality condition (12.). The placenta is more problematic; it has components from both the mother and the foster and so contains both material and fetal DNA. However, as a result of its maternal portion it is strictly topologically connected not to the foster but to the mother, and so it, too, fails to meet condition 12.
12. Stages in the Formation of the Human Substance
We can now give a more thorough account, in ontological terms, of the multi-stage process by means of which the human individual is formed. The single-cell zygote is formed via the unification of two separate substances: the egg-cell and the sperm-nucleus. The parts of the zygote then undergo a duplication within the cell and the topological connections between them are disrupted in such a way that they come to constitute new substances in their own right. These separate substances are now reunified into one at the stage of the early blastocyst. A part of this one individual substance is then detached from the remainder and forms a new individual substance (namely the inner cell mass, or what will become the embryo) in a budding process which involves the simultaneous creation of an interior cavity in which what buds off then resides.
At about the time of the formation of the primitive streak, the embryo undergoes a process of unification or integration. This occurs through the gradual formation of ties between the cells of the aggregate, through the gradual development of a bona fide physical boundary marking out the individual in space, and through the gradual process in virtue of which groups of cells become successively determined to become tissues of specific sorts in reflection of their specific relative locations within the developing individual.
Does it now follow that the human being is formed by degrees across a temporal interval, an outcome which would carry a troubling implication to the effect that different human beings might be human beings to different degrees? To defeat this implication we need to show that the temporal interval across which the mentioned unifying process takes place has some determinate end-point, a bona fide temporal boundary, which we can then identify as the instant at which the human being is formed. The nature of the change which these unifying processes together brings about is to transform what is actually one but potentially many into something that is both actually and potentially one. Consider a collection of viscous bodies floating in a liquid and combining freely together into larger wholes. Suppose that there are chemical processes taking place inside one such viscous body which serve to strengthen its exterior membrane by degrees. At some stage in this gradual process a point will be reached where fusion with other bodies is no longer possible. What is actually a whole but potentially a mere part thereby becomes transformed, abruptly, into something that is a whole both actually and potentially.
We can now begin to understand how a collection of continuous processes can give rise to an abrupt, categorial change. The unification processes which take place in the foster with the formation of the primitive streak together have the consequence that they bring about at some point an integration of the foster that is strong enough to rule out the possibility of twinning. They bring about a transformation from divisibility into non-divisibility. The precise instant of this transformation cannot be ascertained from any clearly visible external mark. Strong evidence for the thesis that the change occurs within or about the period of the formation of the primitive streak is however provided by noting the progressively more serious malformations which arise when twinning occurs within this period. Such deformities, as metallurgists who deal with stress in metals know, are characteristic of abrupt topological change. They give us also extra reason to believe that our account, here, of what it is to be a human being, is more than a mere definitional or conceptual stipulation.
Birnbacher (1995) has objected to such an identification of a time at which the existence of a human individual begins by pointing out that all natural change is continuous; that is, it happens molecule by molecule, and millimeter by millimeter. Thus Birnbacher does not recognize any catastrophes or discontinuities in the realm of natural processes. That Birnbacher is wrong can be seen by moving from the case of temporal discontinuities to their analogues in the spatial realm. We can take objects such as human beings and move them from one place to another without effecting any topological change; but we cannot do this with an entity that is two centimeters beneath the surface of a human being's skin. Human beings have spatial boundaries which are genuine discontinuities even in the face of the continuity of matter in the physical world.
13. The Necessary and Sufficient Conditions for Being a Human Being: An Excursus on Siamese Twins
It would be nice if we could specify not merely sufficient but also necessary conditions for being a human being. Conditions 1.-12. will not serve this end (for individuals of human descent) because of counterexamples in the family of Siamese twins.
Most human beings are bounded spatially, and as in the case of other substances their outer surface provides a bona fide boundary with the external world. Not all entities have bona fide boundaries, however. Mountains, valleys, and body parts have fiat boundaries also. The exceptional cases among humans are provided by Siamese twins.
Siamese twins may result when separation (twinning) occurs after day 10 and succeeds only partially. Siamese twins share as fosters a single placenta, the outer membrane (chorion) and the inner membrane (amnion) and the navel cord. Like identical twins, Siamese twins are the product of a single ovum which splits and forms two centers of organization (two primitive streaks) which lie close together. When these expand and differentiate, they do not divide properly, but fuse and then continue development.
There are many degrees and kinds of connectedness of Siamese twins. There are Siamese twins with only one head. These arise for example, when one parasitic cotwin without a head is attached to a part (the autosite) of the twin, for example to his stomach (heteroadelphi or asymmetrical connection; from the Greek adelphi: brothers). The parasite may even be located inside a cavity in the autosite. This is called 'fetal inclusion', and it usually happens after the fusion of the growing embryos (upon the formation of their respective primitive streaks). One embryo is completely surrounded by the other, and the development of the included embryo is delayed. Heteroadelphi may also result in a tumor-like entity which if not removed is present in the adult, who comprehends a cavity containing a perfectly structured, dead miniature human fetus.
There are other double formations which result in a single individual with only one head which are characterized by a symmetrical development. These are called 'teratoadelphi' (upper connection). The teratoadelphus is typically a Siamese twin with one head and body and four legs some of which can be rudimentary.
It is Siamese twins with two heads and marked by a symmetrical development which are of interest to us here. These involve independent sensory systems which are controlled by their respective brains. The prototype of such Siamese twins are the teratopagi (central connection; from the Greek, pagi: anything that fastens). The twins can here be connected at the stomach or at other parts of the body. They may even be connected at the upper part of the cranium; but they always have two brains which are independent of each other.
Teratodymi (lower connection; from the Greek, dymi: two out of one) have two heads and two spinal cords, but only one body, and one set of legs and arms. They may also have two hearts.
Siamese twins of the latter types are important for our purposes because they constitute one single substance but two relatively isolated causal systems. They are one substance because a part of their external bona fide boundary is completely shared and thus they are strongly connected topologically. There is only one single connected surface, even through some portions of this surface are claimed only by one of the two constituent members of the whole. That they are two causal systems arises in virtue of a partial separation via bona fide boundaries around their respective upper torsos, together with the independent or partially independent circulatory systems.
How, now, are we to amend our twelve conditions in such a way as to provide necessary conditions for human beings which will admit relevant types of Siamese twins? Siamese twins are two distinct human beings if they have two distinct brains and so can respond independently to the environment. Siamese twins are a problematic case exactly for this reason; they are two distinct human individuals and yet they fail to satisfy criterion 7. above. They are therefore more like a limb attached to a body than a substance which has a bona fide boundary of its own. We can, however, amend our twelve conditions by making an exception for human beings who are connected via their outermost parts such as their skin or via organs other than the brain.
The necessary and sufficient conditions for an entity to be a human individual can then be formulated as follows:
x is a human being if and only if
(i) x is of human descent
and either
(ii) x satisfies conditions 1.-12.
or
(iii) x is a part of some whole y which satisfies conditions 1.-9., and x is a relatively isolated system in its own right in the sense that it contains its own mechanisms or procedures to maintain or restore stability, and x is non-divisible (it satisfies condition 11.)
14. How to Refute Our Argument
To refute our argument it would be necessary to show that one or more of the following assumptions is false:
(i) Clauses 1.-12. specify sufficient conditions for being a human being when applied to the products of human reproduction.
(ii) A human being is an entity which maintains its identity across time from the beginning to the end of its existence.
(iii) The beginning of the existence of a human individual constitutes a genuine temporal boundary.
(iv) The foster satisfies condition 12., which means that it is not a part of the mother, but rather related to her as a tenant to its niche.
(v) The single-cell zygote is a substance and a relatively isolated system, and it is maximal, but it cannot be identical with the human being because it ceases to exist through fission almost immediately after it has been formed. Thus it fails condition 11.
(vi) The multi-cell zygote, the morula, the inner cell mass and the epiblast likewise fail condition 11.
(vii) At the end of the formation of the primitive streak, the foster has become a single relatively isolated causal system for which twinning is no longer possible.
(viii) Thus, at the end of the formation of the primitive streak, the foster for the first time satisfies conditions 1.-12. in a way which would make it a candidate to be identical to the human being which exists after birth.
(ix) By the arguments under g.-i. above, there is no later time at which the foster or child undergoes any further substantial change.
Conclusion: The existence of a human individual begins upon the formation of the primitive streak (at around day 16).
References
Aristotle (1984): The Complete Works of Aristotle, J. Barnes (ed.), Princeton: Princeton University Press.
Bermudez, J. L. (1996) "The Moral Significance of Birth", Ethics 106, 2,
Birnbacher, D. (1995) "Gibt es rationale Argumente für ein Abtreibungsverbot", Revue Internationale de Philosophie 3, 357-373.
Brody, B. (1975) Abortion and the Sanctity of Human Life: A Philosophical View, Cambridge, MA: MIT Press.
Casati, R. and Varzi, A C. (1994) Holes and other Superficialities, Cambridge, MA: MIT Press.
Casati, R. and Varzi, A. C. (1999) Parts and Places, Cambridge, MA: MIT Press.
Daly, T. V. (1985) "The Status of Embryonic Human Life - A Critical Issue in Genetic Counselling", in N. Tonti-Filipino (ed.), Health Care Priorities in Australia: Proceedings of the 1985 Annual Conference on Bioethics, St. Vincent's Bioethics Centre, Melbourne.
Diamond, J. J. (1975) "Abortion, Animation and Biological Hominization", Theological Studies 36, 305-24.
Ford, N. M. (1988) When Did I Begin?, Cambridge/New York: Cambridge University Press.
Gaddis, V. and Gaddis, M. (1972) The Curious World of Twins, New York: Hawthorn Books.
Ingarden, R. (1984) Man and Value, Munich: Philosophia Verlag.
Locke, J. (1975) An Essay Concerning Human Understanding (1690), Oxford: Oxford University Press.
Lowe, E. J. (1998) The Possibility of Metaphysics: Substance, Identity, and Time, Oxford: Clarendon Press.
McLaren, A. (1984) "Where to Draw the Line", in Proceedings of the Royal Institution, 56, 101-21.
Novak, M. 1963/64 "A Key to Aristotle's 'Substance'", Philosophy and Phenomenological Research, 24, 1-19.
Simons, P. (1982) "The Formalization of Husserl's Theory of Wholes and Parts", in B. Smith (ed.), Parts and Moments. Studies in Logic and Formal Ontology, Munich: Philosophia, 1982, 113-159.
Smith, B. (1995) "On Drawing Lines on a Map", in A. U. Frank and W. Kuhn (eds.), Spatial Information Theory. A Theoretical Basis for GIS (Lecture Notes in Computer Science 988), Berlin/Heidelberg/New York, etc.: Springer, 1995, 475-484.
Smith, B. (1997) "Boundaries: An Essay in Mereotopology", in L. H. Hahn (ed.), The Philosophy of Roderick Chisholm (Library of Living Philosophers), Chicago and LaSalle: Open Court, 1997, 534-561.
Smith, B. (1997a) "On Substances, Accidents and Universals: In Defence of a Constituent Ontology", Philosophical Papers, 26 (1997), 105-127.
Smith, B. and Varzi, A. C. (1999) "The Niche", Nous 33:2, 214-238.
Smith, B. and Varzi, A. C. (forthcoming) "Fiat and Bona Fide Boundaries", forthcoming in Philosophy and Phenomenological Research.
Smith, B. and Varzi, A. (manuscript) "Parts and Boundaries in Ecology", University at Buffalo.
Spinoza, B. (1677): The Ethics, transl. by R. H. M. Elwes 1955, New York: Dover Inc.
Stjernfelt, F. (1999) "Biosemiotics and Formal Ontology", Semiotica.
Tooley, M. (1983) Abortion and Infanticide, Oxford: Clarendon Press.
1. With thanks to José Bermudez, Peter Hare, John Wilkins and Jiyuan Yu for helpful comments on an earlier version of this essay.
2. We draw here on Aristotle's: Categories, for example, 1b5 and 4a10. Compare Smith 1997a and Novak 1963/64.
3. Substances are not similarly extended in time. Rather, they persist through time, and they exist in toto at every moment at which they exist at all. Thus a substance does not have temporal parts. (The first ten years of John's life are not a part of John, but rather a part of his life, a certain elongated process in which he participates.)
4. What is at issue here is the contrast between rigid and generic dependence (Simons 1982). Each human being is generically dependent on entities of many kinds. Thus a human being must breathe molecules of oxygen; but it is not dependent on any specific molecules of oxygen.
5. This suggests that Aristotle is wrong if he wishes to hold on to the thesis (i) that single-celled organisms are substances, and (2) that no substance has further substances as proper parts.
6. Ingarden 1984, p. 86. The idea of relatively isolated system goes back at least to Spinoza, who, in his Ethics (part 2, propositions 13-16), characterizes a 'union of parts' as being such that it can become greater or smaller and yet retain its nature and its identity. This will be so, Spinoza tells us, if its parts maintain their proportions and mutual communication. Spinoza's preferred example for such a union of parts is the human body. He points out that the human body is hierarchically structured, which means that it contains a number of parts which themselves contain parts, and so forth.
7. Ingarden 1984, p. 90. Apart from the limit case of viruses, 'all life as we know it is defined by membranes, conchs, shells, skins, shields, armor, layers etc. which delimit the internal finely tuned homeostatic system of metabolism from the uncontrollable energy fluxes of the environment.' (Stjernfelt 1999)
9. See Aristotle, Metaphysics, Z3, GA, II, 3, 736a31-b15 and compare Lowe 1998, p. 179.
10. The times indicated here and in the foregoing are of course approximate only. Variations will arise, for example, according to where in the fallopian tube the egg-cell is fertilized.
11. See Locke, Essay XXVII.3. ff. See also Brody 1975, and Tooley 1983 (p. 167) who argues that human life begins when the relevant entity can desire its own continued existence.
12. Another extra-feature view is the development of the soul. Aquinas requires that the fetus, to be a human bein, must possess a human soul. The human soul develops, according to him, gradually and the foster is not a human being during the early stages of pregnancy. According to Aquinas the fetus acquires a human soul at about 40 days for males and 90 days for females. In Petri Lombardi Quartuor Sententiarum: In 3 Sent, dist. 3, q. 5, a.2 co. Compare Summa, I, q. 76, a. 3; q. 118, a.2, ad 2 and II-II, q. 64, a. 1, co.
13. See Gaddis and Gaddis 1972, pp. 30 ff.
14. For the principles governing the tenant-niche relation see Smith and Varzi (1999) and (manuscript).
15. See Smith 1995, 1997, Smith and Varzi 1999.
16. Such strict topological connection is to be distinguished from both weak and strong external connection. Weak external connection obtains, intuitively, when two topologically closed entities touch: the cat and the mat are externally connected when the cat is on the mat. (To say that an object is topologically closed is to say that it includes a complete bona fide exterior boundary as part.) The strong form of external connection arises when one open object, for example a body of air, is in contact with a closed object, for example the walls of a glass container.
17. The foster is in this respect also analogous to a parasite which occupies a host organism. Fosters, like parasites, are functionally dependent on their host organisms; they are both substances and causal systems which are not part of their host organism but are rather related to them as a tenant to a niche. Just as human beings will often react against parasites by producing antibodies, so the mother initially reacts to the presence of the embryo by producing humoral antibodies. But the trophoblast surrounding the early embryo blocks the action of these antibodies so that the embryo is not rejected from the mother's body. One difference between the foster and a parasite in respect to our present argument is that the parasite is not made of human DNA.