Bodily Systems and the Spatial-Functional Structure of the Human Body

Barry Smith, PhD1,2, Igor Papakin1, Katherine Munn1
1Institute for Formal Ontology and Medical Information Science
Faculty of Medicine, University of Leipzig, Leipzig, Germany
2Department of Philosophy, University at Buffalo, Buffalo, NY, USA

Abstract. The human body as conceived by medical science is a system made of systems. The body is divided into bodily systems proper, such as the endocrine and circulatory systems. These are subdivded into many subsystems at a variety of levels whereby all systems and subsystems engage in massive causal interaction with other systems and subsystems. In this essay we offer an explicit definition of bodily system, and explicit means for understanding these causal interactions. Whereas informality is acceptable in documentation created for human beings, it falls short of what is needed for computer representations. In our analysis we will define bodily system, and will take some first steps toward understanding the causal relationships bodily systems have with their subsystems.

1. Introduction

Ontology plays an increasingly significant role in work on terminology and knowledge management systems in the domain of biomedical informatics, and we hold that it will play an essential role in biomedical research of the future. The term ‘ontology’ must, however, be understood in the right way [1]. The dominant paradigm might be referred to as ‘applications ontology.’ This holds that the ontologist should focus primarily on the construction of ontologies as working software applications, a view which often goes hand in hand with the thesis that the expressive power of an ontology is limited effectively to that of one or another version of Description Logic [2]. This means that an ontology, when applied to complex domains such as those of biomedicine, is forced to deal with simplified models. There is, however, a second ‘reference ontology’ school of thought, which focuses primarily on the development of ontological theories of the entities in given domains. Such theories are marked by a high degree of representational adequacy and are designed to be used as controls on the results achieved by working applications rather than as substitutes for these working applications themselves [3]. Three levels of reference ontology can be distinguished in the biomedical domain: 1. formal ontology: a top-level domain-independent theory involving the use of concepts such as: object, process, identity, part, location; 2. domain ontology: a top-level theory applying the structure of a formal ontology to the medical domain, involving concepts such as: body, disease, therapy, organ, tissue, cell; 3. terminology-based ontology: a very large lower-level system, based on medical terminologies such as UMLS, and involving specific concepts such as: inflammatory change in the gastric mucosa. The Institute for Formal Ontology and Medical Information Science in Leipzig is constructing a reference ontology for the domain of biomedicine [4]. This ontology is designed not as a computer application in its own right but as a framework of axioms and definitions relating to such general concepts as: organism, tissue, disease, therapy. Here we focus on the concept bodily system, which we believe will serve as a central factor in a robust ontology of the human organism. Rosse and Mejino [5] have recognized the need in bioinformatics for domain ontologies of the human body, and they and their co-workers are creating such a domain ontology of human anatomy, the Foundational Model of Anatomy (FMA). The FMA symbolically represents the structural organization of the human body from the macromolecular to the macroscopic levels, with the goal of providing a robust and consistent scheme for classifying anatomical entities on the basis of explicit definitions. This scheme also provides a template for modeling pathology, physiological function and genotype-phenotype correlations, and it can thus serve as a reference ontology in biomedical informatics. Our aim, like that of Rosse and Mejino, is not applications ontology, but rather a reference ontology. That is, we seek to provide a theoretical framework that, when fully developed, can serve as the theoretical basis for applications ontologies. Whereas Rosse and Mejino’s ontology represents the anatomy of the human body in the purely structural context, our ontology of bodily systems will encompass anatomical structures in a context which takes account of functions and their realizations. We aim, in other words, to link anatomy with physiology. One way in which our work supplements that of Rosse is in providing a definition of the notion of ‘system.’ The definitions of the FMA use this term, but leave it unanalyzed. We believe that the concept of bodily system can serve as a unifying top-level element in a robust ontology of the human organism designed for the purpose of medical informatics. Concepts like system and function are used throughout contemporary medical science, and medical references and textbooks overflow with reasoning about systems of different kinds which are constituents of the body. Much implicit medical knowledge factors into such representations. Our goal here is to apply philosophical rigor to a definition of ‘bodily system’ in a way that makes this knowledge explicit.

2. Bodily Systems in Medicine

Contemporary medical science represents the living human body, or the human organism, as a system made of systems. The body’s systems serve as major provinces in our maps of human anatomy; they thus play a central role also in a variety of domains, from medical pedagogy and dynamic modeling to computer visualization. An understanding of the concept system is moreover a necessary part of any understanding of cognate concepts such as organ and function, and it is a prerequisite for understanding systemic diseases, both those which are localized in single systems and those, such as diabetes, which affect a plurality of systems simultaneously. We are interested primarily in the top-most nodes in the hierarchy of bodily systems; i.e., in those major bodily systems towards whose functioning other, smaller systems contribute. In Table 1 we provide examples of divisions of the human body in terms of such highest level systems derived from standard sources. Clearly, the division of the body into its major systems is by no means unproblematic. Medical textbooks rest on informal explications of the systems that concern us here, and rarely if ever do they offer anything more than a passing description of the concepts that these explications presuppose – including the concepts of ‘function’ and ‘system’ themselves. While such informality is acceptable in documentation created for human beings, who can use their tacit knowledge of the entities involved to achieve sufficient understanding, biomedical information systems will require precise and explicit definitions of the relevant terms. The analysis presented here is intended as a first step towards providing a framework for such definitions.

|Terminologia anatomica [6] |National Library of Medicine |Wolf-Heidegger’s Atlas of Human Anatomy |Foundational Model of Anatomy [5] |
| |[7] |[8] | |
|Bones, |Musculoskeletal system |Motor system |Body |
|Skeletal system; |Respiratory system |Bones |Integumentary system |
|Joints, |Cardiovascular system |Joints |Musculoskeletal system |
|Articular system; |Hemic and lymphatic |Synovial joint, muscle, tendon |Nervous system |
|Muscles, |systems |Skin |Neuraxis |
|Muscular system; |Gastrointestinal system |Skin and fingernails |Peripheral nervous |
|Alimentary system; |Urogenital system |Circulatory system |system |
|Respiratory system; |Endocrine system |Cardiovascular system |Autonomic nervous |
|Thoracic cavity, |Nervous system |Adult circulatory system |system |
|Thorax; | |Fetal circulatory system |Somatic nervous system |
|Urinary system; | |Blood vessels of trunk |Enteric nervous system |
|Genital system; | |Lymphatic and organ systems |Sympathetic nervous |
|Abdominopelvic | |Lymphatic system and endocrine organs |system |
|cavity; | |Digestive and respiratory |Parasympathetic |
|Endocrine glands; | |systems |nervous system |
|Cardiovascular | |Urogenital system |Hematopoietic system |
|system; | |Central and peripheral nervous |Cardiovascular system |
|Lymphoid system; | |systems |Alimentary system |
|Nervous system; | |Central and peripheral nervous |Urinary system |
|Sense organs; | |systems |Male genital system |
|the integument | |Spinal cord and spinal nerves |Female genital system |
| | |Automatic nervous system |Respiratory system |
| | |Vegetative nervous system |Hemolymphoid system |
| | | |Endocrine system |

Table 1: Brief overview of bodily systems distilled from four standard sources

2.1 Analyzing Traditional Views of Bodily Systems

Partitioning the body into systems in the way that we do is a cognitive process. But it is a cognitive process which involves representing aspects of reality: the body is not a continuum, but rather is already partitioned on various levels into constituents of various sorts. This is why, in constructing a domain ontology of medicine, we need to start with the results of medical science as set forth in standard textbooks. Unfortunately the medical literature provides at best informal definitions of terms such as ‘bodily system,’ ‘organ,’ and so forth, which helps to explain why there is a less than perfect agreement on the rosters of bodily systems provided by different sources. Medical textbooks rest on tacit knowledge concerning such highly general concepts. That is, while their authors understand perfectly well how the living human body is organized and what the functions of bodily systems are – they deal with such systems and their workings every day of their lives – they do not formulate this knowledge in an explicit way. We do not wish to impose spurious precision in an area that is marked intrinsically by a certain vagueness and informality. But again: where for human beings a high degree of informality is acceptable, it becomes problematic where human reasoning must be simulated inside a computer. The analyses presented below are intended as a first step towards filling this gap. Distilling an overview of the adult human body’s repertoire of systems from the sources listed on Table 1 gives us the following [8]:

Supportive Systems: These provide the framework, or structure, within which the other systems are located. They include: The skin system separates and isolates the organism from its surroundings and participates actively in maintaining the organism’s internal environment. Its functions include thermoregulation, tactile sensitivity, participation in the maintenance of water balance, and defense (it has an innate and acquired immune response against bacteria and viruses). The musculoskeletal system is an ordered assembly of bones, muscles, and ligaments that maintains the body’s shape and allows movement in counteraction with external forces such as gravity. It also creates an internal framework of support for the organs of the body.

Systems for Substance-Exchange: These support the normal ways in which the organism exchanges substances both within itself and with the surrounding world. They include: The digestive system, which itself includes inter alia the oral cavity, the esophagus, stomach, duodenum, small and large intestines (the digestive tract), and additional parenchimatous organs like the salivary glands, liver, and pancreas. This system ensures that solid and liquid substances are absorbed into the body in such a way that they can serve as a source of energy and as building blocks for other bodily systems. The respiratory system includes the nasal cavity, larynx, lungs, trachea, and bronchus. It performs respiration, i.e. it enables the absorption of oxygen and the excretion of carbon dioxide to take place within the body. The circulatory system includes the heart, blood vessels, and microcirculatory vessels, as well as the blood itself. It serves as a universal transporter of substances to the body’s cells via two circuits: the pulmonary and the systemic. The pulmonary circuit exchanges gases with the external medium in the lungs. The systemic circuit supplies all the organs of the body with oxygenated blood and provides for gas exchange between the blood and the cells of other organs. The urogenital system includes the kidneys, the ureters, the urinary bladder, and the urethra, as well as the external and internal genitals (reflecting the close anatomical and functional relation of the latter with the urinary organs). It is responsible for the excretion of surplus water, and of the waste products that appear in the cells as a result of physiological processes, and for regulating the body’s ion balance.

Systems for Regulation: These act as supervisors and coordinators in the work of the other bodily systems: The immune system includes the thymus, bone marrow, spleen, lymphatic nodules and lymphatic vessels, as well as the lymphatic tissue in the pharynx, the intestine, and the population of immunocompetent cells working through the body. Its first task is to recognize and break down or eliminate substances that might damage the body’s integrity. The nervous system includes the brain and spinal cord (the central nervous system), together with the peripheral nerves, ganglia, plexi and sensory organs. It regulates all the body’s systems and provides the sensitive (sensory) functions of the body. The endocrine system consists of the endocrine glands and of the active endocrine tissue in other glands. Together these serve as a battery of transmitters that broadcast instructions to all the cells of the body. Only those cells that have a ‘receiver’ – i.e. a specific receptor in their membranes – can recognize and use this regulatory sign.

The above is provided merely as a preliminary orientation for the discussions which follow, for the division of the body into its major systems is by no means clear-cut. There are elements of exchange systems (for example parts of the liver and pancreas) which play a role also in regulatory systems, and the three regulatory systems themselves realize their functions of regulation via a certain sort of substance-exchange. The rationale for our proposed division will, however, become clear in the course of what follows.

3. Toward an Ontology

The task of reference ontology is not to replace medical science. Rather, its job is to provide a framework within which medical knowledge can be formalized in a way that supports causally predictive theories, and at the same time counteracts the effects of terminological and other inconsistency and imprecision. Such a framework must start out from the ways knowledge is formulated in the medical literature, and one criterion of a good definition of ‘system’ is that it yields a roster of systems that is very like the standard rosters. As we have seen, however, it needs to go beyond textbook formulations if it is to achieve the sort of formal clarity we need for the purposes of reference ontology. How, then, are we to define the notion of a bodily system? The discipline of systems theory is of little help to us here, since its definition of a system as a complex of interacting parts [9] is far too general for our purposes and is made more specific only by the use of mathematical tools which leave unanswered precisely those questions pertaining to the nature of bodily systems which we are called upon to answer. We can make some progress, on the other hand, if we examine how the word ‘system’ is most commonly used in both technical and non-technical contexts by speakers of English. The Oxford English Dictionary defines ‘system’ under the principal heading of ‘an organized or connected group of objects,’ or more precisely: ‘A set or assemblage of things connected, associated, or interdependent, so as to form a complex unity; a whole composed of parts in orderly arrangement according to some scheme or plan.’ Under the heading ‘Biology’ it gives: ‘A set of organs or parts in an animal body of the same or similar structure, or subserving the same function, as the nervous, muscular, osseous, etc. systems, the digestive, respiratory, reproductive, etc. systems.’ One might be critical of such definitions on the grounds that a system is not a mere set or aggregate but rather something dynamic (think of the solar system). We can do justice to such criticisms, however, by distinguishing systems themselves from the processes in which they are involved, or in other words from the functioning of systems [10]. As we shall see, systems are able to carry out processes, or to realize their functions, only via a physical structure, or aggregate of parts. Systems on this view are no more dynamic in nature than organisms. Indeed organisms are systems on the analysis we shall defend. Examining our list of systems above, we see that each of them consists of a certain organized or somehow connected group of objects – such as organs, associated tissues, and populations of cells – to which some complex function is attributed. Unfortunately there are many organized collections of bodily parts with which functions can be associated, including each and every individual cell. So in order to make an analysis of ‘system’ along these lines work, and yield the sorts of answers we need for our questions about bodily systems, we will first need to provide a specification of the particular notion of ‘function’ that is at issue here. And before we do this we must discuss the organized parts of the body that have functions.

3.1 The Body’s Modular Hierarchy

There is a collection of bits of biological matter in the human body that medical science designates as the circulatory system. What is it about this collection of vessels, organs, and blood in virtue of which it is referred to as a system? Could we designate as a system our heart, plus our salivary glands, plus our big toenail, plus our right ear? The reason that the former are considered a system and the latter a mere conglomeration is because the heart, vessels, and blood are related to each other in special ways. There are anatomical connections, which can be described in terms of mereological relations: the left ventricle is part of the heart, a capillary is attached to a venule. And there are physiological connections, which can be described in terms of causal relations: the myocardium causes the heart to pump by electro-chemical stimulation. In other words, a system is characterized simultaneously by a certain complex structure, and by a set of processes in which that structure participates. It is this complex structure that allows the processes to participate as they do. Without a tendon connecting a muscle to a bone and a group of motoneurons connecting the muscle to the central nervous system, the process of movement could not take place. Because there is such a multiplicity of complicated processes that takes place in the body, the structure that accommodates them is also complicated. This structure is a modular one: the body as a whole can be divided into units, each unit can be divided into sub-units, and so on. Every complex organism has modular units at many levels. Your brain contains neurons, the neurons contain organelles, the organelles contain molecules, which are composed of atoms, which are composed in turn of subatomic particles. Modular units at lower levels have mereological relationships vertically with modular units of which they are the parts. The alveoli are lower-level parts of the respiratory system, and are parts of the lungs, which are higher-level parts of this system. Modular units also have causal relations horizontally with modular units in other systems, and vertically with modular units of which they form a part. The alveoli are parts of the lungs and have a function in the context of the respiratory system. The alveoli have a horizontal relationship with the capillaries, which have a function in the context of the circulatory system. It is the alveoli and the capillary wall where the exchange of oxygen and carbon dioxide takes place between the blood and the air. Modular units also have diagonal causal relationships with units from other systems. The liver is for example an element in the alimentary system, but it has functions in the contexts of several other bodily systems at various levels. In the context of the circulatory system, it produces proteins for the coagulation of blood. In the context of the digestive system, it produces bile for breaking down chyme. Bile also has a function in the context of the excretory system: the body mixes waste with bile and excretes this mixture as feces. The liver also produces proteins that have a receptor function in the context of the immune system. It is this specific and complex structure that allows physiological processes to be carried out within the body. This structure enables certain items in the body to causally relatively isolated from other items. This causal relative isolation is what enables processes to take place without getting in the way of other processes. It is also what allows certain processes to have causal relationships with certain other processes. As the philosopher Roman Ingarden expressed the matter, each multi-cellular organism is a relatively isolated system of a very high order, and as such contains in itself very numerous, likewise relatively isolated, systems of lower and lower levels, which are hierarchically ordered and variously situated within the organism, and are at the same time both partially interconnected and also partially segregated, as a consequence of which they can exercise the specific functions which are characteristic to them relatively undisturbed [11].
Our task here is to provide the beginnings of an account of this modular hierarchy and of the layers from which it is built, from macromolecules via cells and organs through to the whole organism, but in a context in which we pay attention to the functionings of the various modular units. To anticipate somewhat, we can say that the highest level of this modular organization is the whole organism itself and that the level immediately beneath is that of bodily systems.

3.2 SNAP and SPAN in Bodily Systems

Ontology offers tools for formalizing the aspects of anatomy and physiology we have just discussed. The first tool we will need provides us a way of distinguishing between the body’s complex structure and the processes in which that structure participates. The structure can be understood as occupying three dimensions of space; the processes require in addition the four dimensions of time. Ontology also provides a way of talking about the relationship between the structures and processes. What is called for is an ontology that distinguishes between three-dimensional objects that endure through time (endurants, for short), and the four-dimensional processes (perdurants, for short) in which these objects participate. Grenon and Smith provide such an ontology [12]. The body’s structure is three-dimensional: it can be apprehended as it were in one glance as in a snapshot; hence Grenon and Smith call it a SNAP ontology. A SNAP ontology of the circulatory system includes endurants such as the heart and the blood. The processes that take place in the body are four-dimensional: they cannot be captured in a snapshot, but require instead something like a videotape, which allows them to be captured in their temporal extendedness as they unfold over a certain timespan – hence Grenon and Smith call it a SPAN ontology. A SPAN ontology of the circulatory system includes perdurants such as the beating of the heart and the flowing of the blood. In a SNAP ontology, endurants are visible but perdurants are not; in a SPAN ontology, perdurants are visible but endurants are not. SNAP and SPAN thus represent two complementary perspectives on the same reality. In order to talk about bodily systems, we need both of these perspectives. We need a SNAP ontology of the endurant parts of the body that make up the modular hierarchy, and we need a SPAN ontology of the perdurant processes that this structure allows for. This ontological distinction between endurants and perdurants helps us explain why the heart, blood, and blood vessels comprise the circulatory system, whereas our heart taken together with our salivary glands and our right ear comprise no system. Endurants in the body must have certain causal relationships with each other in order to constitute a system in the sense that is relevant to us here. The heart, blood, and blood vessels are parts of the human body. They are also parts of the circulatory system. A SNAP ontology suffices to show us how these parts relate to each other spatially. But in order to see how these relations play out in the form of processes in which the corresponding objects participate, we need a SPAN view. We must distinguish, then, between two kinds of parts: parts of the body, which can be seen in SNAP alone; and what we will call elements of the body, which are the results of demarcating the body into systems in a way which involves taking account of not only the SNAP but also of the SPAN ontology. The concept of element can be understood, roughly, as a generalization of concepts such as organ, cell, or even bodily system. Whereas parts are the building blocks from which the body is constructed, elements are the building blocks from which bodily systems are constructed. At the highest level of the modular hierarchy, the bodily systems proper are elements of the system that is the whole human body. If it turns out that there are nine top-level bodily systems, then the human body will be a system composed of nine elements. It almost goes without saying that all elements of the body are also parts of the body. Not all mereological parts, however, are also elements. The heart is an element of the circulatory system, and it is a part of the circulatory system’s complex physical structure, which is in turn a part of the body. The foot is a part of the body, but is an element of no system. Certain parts of the foot, however, such as its bones, capillaries, or nerve endings, are elements in various systems. The bones are elements in the skeletal system, the nerve endings in the nervous system, and the capillaries in the circulatory system. Elements are distinguished by the fact that they have special kinds of causal relations with larger wholes, called systems. If an element becomes causally disconnected from its system, as when a heart is refrigerated in the course of a heart-transplant operation, then it ceases to be an element for a certain period of time. As Aristotle expressed it: ‘A dead body has exactly the same configuration as a living one; but for all that is not a man…. no part of a dead body, such I mean as its eye or its hand, is really an eye or a hand’ [13]. While the elements in a given system are distinguished from parts in general within the system by their causal connection with other elements, they are distinguished from other systems by their causal relative isolation from their surroundings. This causal relative isolation at many different levels is part and parcel of the body’s complex structure, and it is what allows for elements to have functions within the workings of the body as a whole. Some examples of elements and their corresponding functions: The elements of the digestive system include the esophagus and the stomach, the serous membrane, the layers of smooth muscular tissue, and so forth. Some corresponding functions are: to provide the way for a bolus to pass from the mouth cavity into the stomach, to advance the process of digestion by mixing the bolus with hydrochloric acid and pepsin (stomach juice), to allow for the external coverage of the stomach and its constriction respectively. The heart is a system consisting of the myocardium, endocardium, and so forth. These elements have their own specific functions and comprehend their own elements in still another layer (for example different types of cells). The liver consists of several types of tissue, including the hepatic parenchyma, connective tissue, and capillary vessels. The blood consists of cellular elements (red blood cells, leucocytes, lymphocytes, monocytes, platelets) and plasma, which contains albumins, globulins and hormones. Each cell is a system that consists of elements such as nucleus, mitochondria, endoplasmic reticulum, ribosomes, which are in turn systems in their own right with their own specific functions. As we shall see, elements are often shared by different systems, and they then manifest distinct functions – one or more in the context of each system to which they belong. For instance, the pharynx has the function of enabling both the passage of the bolus into the digestive tract and the passage of air into the lungs; as such it is an element of the digestive and respiratory systems simultaneously.

3.3 Granular Partitions and System Elements

The second ontological tool we will need is a theory of granular partitions [14]. This provides a way of formalizing our description the structure of the body’s modular hierarchy. A theory of granular partitions represents reality in terms of partitions each of which highlights entities of a different grain. An organism is a bona fide object: it exists independently of our partitions. An organism is also a totality of atoms. At the same time, it is a totality of cells, and a totality of molecules. It is also a single unitary substance. All of these express distinct granular partitions of the organism. A theory of granularity preserves realism, even as it accounts for the possibility of having different perspectives on reality. Each granular partition highlights certain aspects of the unified whole. Think of a tourist map as a granular partition: it represents a given city, but it highlights only certain selected tourist locations. A map of bus routes is another granular partition of the same city. Each grain in the partition is an item on the map. Grains themselves are divisible into ever smaller grains. All of the items appearing on the tourist map have one attribute in common: they are of interest to tourists. All of the items on the bus map have something else in common: their relevance to the purpose of navigating through the city by bus. Similarly, a granular partition of bodily systems highlights those items in the body that have some feature or features in common, and leaves out items that do not have these features. What is the feature in virtue of which a granular partition includes bodily systems? Recall that a system is characterized simultaneously by a complex modular structure, which is a SNAP entity, and by a family of associated processes, which are SPAN entities. The processes occur as they do because the body is structured in the way that it is. Thus the processes that this structure allows for are highly specific. These structures and processes, then, are the features that granular partitions of bodily systems and the associated bodily processes must highlight. Your right leg from the knee down would not appear as a modular unit in a granular partition of bodily systems: there is no physiological process in your body that is dependent on its being structured in this way. The most, then, that we can say about your right leg from the knee down is that it is a part of your leg, which is a part of your body. A granular partition of human anatomy might include your right leg from the knee down (it is an interesting question which connected parts of the body ae properly included in a strictly structurally-based ontology of human anatomy [15]); but in a granular partition of bodily systems it would be invisible. Your heart, on the other hand, is a structure on which the carrying out of certain physiological processes depends. It is one link in a causal chain that results in blood being pumped through the veins. Your heart, then, is visible in a granular partition of bodily systems, and so also are your veins. This is because heart and veins are modules in the hierarchy that enables certain associated physiological processes: they are elements of the circulatory system. Further, your aorta and ventricles appear in the same granular partition, because they are crucial to your heart’s structure, and to the processes the structure enables; they are also elements of your circulatory system.

3.4 Functions in Bodily Systems
A granular partition that is subdivided into many different grains is needed to highlight the human body’s systems and their elements on successive levels. Each element is distinguished by a specific structure that allows for specific physiological processes. In the everyday language of life science, this element is said to have a function. A function, like an element, is a SNAP entity or endurant. Unlike an element, however, a function is an endurant that is ontologically dependent on another endurant. For example, the heart has the function: ‘to pump blood’ – this function could not exist if the heart did not exist. Similarly, color is a dependent SNAP entity: your eyes have a token color (brown) which would not exist if your eyes did not exist. A function is dependent for its existence on a SNAP entity; but it is realized in a process, a SPAN entity. The heart’s function: to pump blood is realized in the process of blood being pumped. It may now look as though recognizing both functions and processes in an ontology represents a case of ontological double-counting. Certainly it is true that every function is correlated with some class of processes (of functioning) at some given level. But many processes are not realizations of functions, including for example all those processes in the human body which are malfunctionings. Also, many elements have functions which are overwhelmingly never realized in processes. It is not only true of, but essential to the nature of, fish eggs that, other things being equal, they develop into fish. The statistical situation is even worse than usual: in many species, the vast majority of eggs never develop into fish, most being eaten or destroyed. And yet each fish egg has the function to develop into a fish, not, say, the function to be eaten by predators in order to protect the other eggs. It is a contingent fact that most fish eggs do not develop into fish. Thus the concept of function cannot be reduced to explanation in terms of processes. A quick word about the concept of function in philosophy. Some philosophers have criticized the function-talk that life scientists use unreflectively (for example in the designation of disciplines like functional genomics), and have tried to reduce the notion of function to the notions of causality or natural selection.[1] We hold, however, that biological functions are real, irreducible features of the world. With Johansson et al. [19], we support the view that the functions of bodily systems and their elements are part-whole functions: that is, functions of parts within the context of a larger whole. This follows from our understanding of elements as components of a bodily system distinguishable by their structure and by the specific processes which that structure enables. An element is only an element of some larger system. Thus the function that the element has can also only exist within the context of this larger system. Further, it is only in the context of a larger system that the function can be realized in a process of functioning. The resulting picture is a complex hierarchy, or taxonomy, that is at once spatial and functional. The heart, for example, is at once a part of the circulatory system and an element in that system. As a part, it is a mereological component of a physical structure visible in a SNAP ontology. As an element, it has a function that is realized in processes, and therefore requires reference also to a SPAN ontology. On the spatial-functional hierarchy proposed by Johansson et al., the circulatory system is at the top level, the heart is located at the next level down, and its elements: ventricles, atria, valves, and so on, at the next level thereafter. Each of the latter bears a function in relation to the higher-level functioning of the heart. On the account we will defend, the circulatory system itself is an element of the human body proper. Now, of course, we run into a problem: what is the function of the human body, in relation to which the circulatory system and other bodily systems have part-whole functions? Johansson et al. address this problem by distinguishing between several possible types of functions the human body may have: (a) a part-whole function of some larger whole, say a species; (b) an objectively existing function, which is intrinsic to the human body, or (c) a function purpose merely ascribed or imputed to the human body in the conventions of language-using human beings along the lines proposed by John Searle [16]. Picking (c) is inconsistent with our presupposition that the functions of the bodily systems exist in objective reality. It would also, as Johansson et al. note, ‘seem to license cultural relativity or even pure subjectivity to enter into science’ [19, p. 3]. Science itself, on the other hand, often talks about functions intuitively as if they exist in objective reality. Johansson et al. point out that neither (a) nor (b) contradicts the proposition that part-whole functions exist objectively, and it is possible to hold either (a) or (b) consistently with an account of part-whole functions at every level below the human body itself. Even so, both (a) and (b) are problematic for their own reasons: (a) is problematic because it calls for an account of what the larger whole is within which the human body functions, and further seems to be thereafter a vicious regress: where does the spatial-functional hierarchy stop? (b) is problematic because there is as yet no successful scientific definition of ‘intrinsic function.’[2] Here, therefore, following Johansson et al., we bracket the question of what kind of function the human body has, and concern ourselves only with the functioning of its elements.

3.5 Spatial-functional Hierarchy

We are now prepared to go into further detail about the body’s modular hierarchy. What we have said so far is enough to show that the concept of modularity is too narrow to describe bodily systems. While modularity does describe the structure of the human body, this is all it describes. A hierarchy of systems must encompass physiology as well. To this end Johansson et al. propose a spatial-functional hierarchy, which yields a taxonomy of the body’s anatomical structures in addition to the functions these structures have. As we will see, the body’s anatomy and physiology lend themselves readily to such a taxonomy. The spatial half of the hierarchy taxonomizes the body’s anatomy according to a modular structure (i.e. in terms of what is a part of what): a nephron is a part of a kidney. The functional half of the hierarchy taxonomizes the body according to which functions cause, or enable, which other functions to occur. One nephron has the function of removing waste from the blood, which is the function that the kidney as a whole performs on a large scale. Fusing a spatial taxonomy with a functional taxonomy yields a spatial-functional hierarchy. As we might suspect, the two halves match up. Why would we suspect this? Because, as we mentioned earlier, functions are dependent SNAP entities which depend for their existence on independent SNAP entities. The body’s complex anatomical structure allows for processes to occur, which means that the body is structured in such a way that the functions realized by the body can be ordered in a hierarchy which parallels the structure. Thus a cellular mitochondria in a myocyte provides the chemical energy without which the myocyte cannot contract, and therefore contributes to the pumping of the heart.

4. ‘Element’ Defined

With an ontology that can account for endurants and processes, with a theory of granular partitions, and with a spatial-functional hierarchy of part-whole functions, we are now prepared to define ‘element:’

X is an element of Y if and only if:

(i) X is lower on the spatial-functional hierarchy than the organism as a whole; (ii) X has one or more specific functions; (iii) X is causally relatively isolated from the parts of the organism that surrounds it; (iv) X is maximal, in the sense that an element is not a proper part of any item on the same level of the spatial-functional hierarchy satisfying conditions (i) to (iii).

Ad (i): An element is an element only in the context of a given system, within which it has a part-whole function. The body as a whole is not an element of any larger organic system. Thus only items that are proper parts of the body can be elements. Ad (ii): Causal relations are made possible by the complex structures of the system and its component elements. In virtue of this complex structure, and the processes which that structure enables within the context of its overarching system, an element has a function. Functions in bodily systems are in every case part-whole functions relative to some larger whole. It should be noted that there is no reason to exclude an element’s having more than one function in the context either of the same system or of different systems. The liver has relative to the digestive system the function: to produce bile. Relative to the circulatory system it has the function: to produce plasma enzymes that contribute to the clotting function of the blood. A function is a dependent endurant entity; it is the function of some element (like the heart). It realizes itself in perdurant entities, also called processes or activities. However, it must be noted that in order to exist a function need not be realized at all times, or even once in the course of an organism’s lifetime. For example, whether or not a woman becomes pregnant, a function of her uterus is to make it possible to house an embryo. This is the function of her uterus even if for some reason she cannot have children. Further, according to our definition of element, only those entities which have part-whole functions in the body are elements of the body. Thus a virus may take on a functional role in your body, directing the cell to construct certain proteins that the virus needs for reproduction. The virus is however not an element of your body – indeed it is not even a part of your body – because the directions given by the virus interfere with your body’s functioning [21]. Ad (iii): The complex structure of bodily systems and elements enables processes to take place by providing elements which enjoy a relative causal isolation from other elements in the same and other systems. The causal processes mentioned in (ii) can only occur if other causal processes do not interfere. In other words, at least in the sorts of cases that concern us here, tight causal connections within and between elements require some degree of causal isolation from the processes of other elements. Each element is partially isolated from certain outside causal influences (for example by the presence of a porous membrane which at the same time allows certain kinds of influences and substances to encroach into its interior). This relative causal isolation is what allows systems or elements to be self-contained yet responsive to stimuli from the outside. It is the complex structure of the body, which allows for the body’s tight causal organization, that enables body systems and their elements to engage in causal relationships with their environments (which include other bodily systems and elements), yet not have their integrity threatened, or break down, as a result of these relationships. Ad (iv): The spatial-functional hierarchy of the human body works on the supposition that every element is composed of elements that enable it to realize its function, and that this is the case all the way down to an as yet unspecified bottom level. Elements are maximal: this means that on any given level of the hierarchy an element is not a proper part of any element on the same level. A cell is an element; a half-cell is not. The fibula is an element; the upper fibula is not, because it is not maximal. The principle of maximality is linked to the spatial-functional hierarchy via the connection between function and relative causal isolation. An element’s relative causal isolation is what enables it to engage in substance exchange; this substance exchange is in turn what produces the function. Thus physical maximality as we understand it here entails causal relative isolation. The bottom half of a lung, for example, does not have a separate function from the whole lung, for the lung is physically maximal. As such it is causally relatively isolated from the the rest of the organs and fluids in the thorax. The lung’s relative causal isolation is what enables it to exchange oxygen and carbon dioxide without disrupting or being disrupted by other processes in the body. If this causal relative isolation were disturbed, it would no longer be able to realize this function – and other elements in the body would be prevented from realizing theirs. Rupturing a lung would cause air to flow from it to the thorax. Not only would the lung be prevented from realizing its function effectively, many of the elements in the thorax would become too compressed to be able to realize theirs. Just as systems can be divided into elements, so functions, too, can be divided into sub-functions (corresponding to the elements which perform them). Functions located at lower levels of the spatial-functional hierarchy interact in complex ways to enable functions at higher levels. For example, the function of a particular neuron, e.g. to provide a path for electric impulses, is realized in a composite process that consists of smaller interrelated processes, such as the exchange of potassium and sodium ions through the cellular membrane. The function of the kidney is to excrete urine. This function is executed via a composite process that consists of smaller interrelated processes that occur on lower levels of granularity: the excretion of urea and creatinine, absorption of necessary ions and excretion of redundant ions and water.

5. Demarcating Bodily Systems

The relation between elements and functions is complicated by the fact that there is no perfect one-to-one correspondence between them. This is because many elements in the human body have a multiplicity of functions. Recall that the body’s elements are unified together within a single whole – namely the corresponding bodily system – which is able to regulate its own state and structure, and that it is in the context of this whole that elements have the functions they have. On this account, the blood has several different functions in several different systems. In the context of the digestive system, the blood’s function is to transport nutrients and allow for nutrient and waste exchange at the cellular level; in the context of the respiratory system, its function is to transport gases and allow for gas exchange at the cellular level. Blood, therefore, like most elements, can be located simultaneously at different horizontal levels of the spatial-functional hierarchy, for it has different functions within the context of each system. We must now discuss the interconnections among the bodily systems in more detail. We have thus far pointed out merely that bodily systems and their elements are distinguished first by a complex structure that enables causal connection and causal isolation of elements. We have also pointed out that elements have functions relative to the larger system to which the elements belong, and that the largest system that is a whole in a part-whole function is the human body itself.

5.1 Evaluating Functionings

Medical science can demarcate one bodily system from another in terms of the way each contributes to the task of keeping the human body alive. It now becomes possible to evaluate a particular functioning of a particular type of system based on the success or failure of its contribution to this functioning of the whole. The spatial-functional hierarchy gives us objective criteria in terms of which we can effect an evaluation of functionings. In a spatial-functional hierarchy built from a taxonomy of part-whole functions, an element succeeds at performing its function when that performance contributes to the whole in such a way that the whole can contribute to the functioning of the larger whole of which it is an element. And so on, until we reach the maximal whole, the human body. The body’s survival becomes the benchmark of evaluations of functionings. For example, all else being equal, a circulatory system with clogged arteries is less efficient, and therefore less successful, than a circulatory system with clear arteries. This is because the former contributes poorly to the survival of the body as a whole relative to the latter. There are in the real world degrees of functioning in each case relative to one or more prototypical functionings. Johansson [22] points out that an element’s functioning can be measured by a prototype in a similar way to that in which an object’s weight can be measured by a scale. In order for a functioning to be successful, it need not match the prototype. Thus a screwdriver can still realize its function even though its head is somewhat loose. There are many dimensions along which a functioning can be plotted in relation to its prototype. One dimension we are concerned with here is that of the success or failure of a given functioning to enable functionings at higher levels. Presumably there is an ideal, or prototypical, pumping of the heart that contributes optimally to the survival of the body as a whole. If the pumping of the heart is disabled by a myocardial infarction, then its functioning can move far enough away from the prototypical functioning that it no longer succeeds in supplying the brain with oxygen. As we move away from the prototype, we can order actual pumpings, and actual transport of blood by the arteries, according to the degree to which they contribute to the systems that their functions are parts of. Once a functioning crosses a particular threshold at a particular distance from the prototype, it is no longer successful enough to play its part in the functioning of the whole system that it contributes to, in ways which must lead either to death or substitution. The threshold for success of a given type of functioning – for example, the taking in of oxygen by the lungs – is relative both to the individual organism and to its specific environment at some given time. A bioinformatician sitting at a computer all day has a threshold for successful oxygen intake that is much farther from the prototype than a manual laborer on a high-altitude farm in the Andes. The survival of the body as a whole can be used as an objective standard for evaluating functionings in the body according to their proximity to the prototype functioning that is most conducive to the functioning of the system one level up. As we will see, having this objective standard of evaluation brings us one step closer to showing the reasoning behind medical science’s demarcation of bodily systems.

5.2 Critical Functions

We have just pointed out that functionings can be ordered according to their success or failure in contributing to the functioning of the overarching system to which they belong. A functioning is successful if it matches or comes close to the prototype, and it fails the instant it crosses a threshold beyond which it no longer contributes sufficiently to the functioning of the higher-level system. It is also possible to order functions themselves. There are many ways in which functions can be ordered. The way that concerns us here is to order functions on the basis of the degree of importance they have for their functional whole. The heart’s function to pump blood is for example more important for the survival of the body than is the function of one skin cell to guard the body against the invasion of foreign substances. Or, as we will say, the heart’s function is more critical to the survival of the body than is the function of one skin cell. After we subject the notion of criticality to further analysis, we will see that it is this dimension of evaluation of functions that yields the principle for the division of the body into its major systems. But first a note about the difference between a successful functioning, which is a SPAN entity, and a critical function, which is a SNAP entity. We have suggested a means of evaluating functionings qualitatively, based on their proximity to a prototype functioning. The latter exemplifies the ideal functioning of an element in relation to the ideal functioning of the system it belongs to. Evaluating functionings means comparing one type of functioning, say the heart’s beating, with the prototypical functioning of the ideal heart. Ordering functions according to how critical they are to the body’s survival, on the other hand, means comparing one type of function with another type of function, say the heart’s function to beat with a skin cell’s function to protect the body from foreign invasion. The terms of comparison are: how critical is a given type of function to the function of the body as a whole? It should be noted that we use the notion of criticality in a somewhat non-standard sense; or rather, that we are expanding on one standard sense of the term. In everyday speech, criticality refers to some highest degree of importance: a drought can approach criticality by becoming more and more severe. In physics a point of criticality is the point at which a nuclear reaction becomes self-sustaining. In medical science, criticality refers to the point at which the body can no longer survive: a disease is critical if it threatens the patient’s life. Along similar lines, we understand a function to be critical if the body as a whole cannot survive without it.

F is a critical function for organism Y if and only if [23]:

(i) some element of Y has F as its function; (ii) the continuing to function of organism Y is causally dependent on the continued performing of F by X.

The function of the digestive system is critical to the survival of the body, as is that of the immune system. As we will see, then, criticality can also admit of degree. The function of the vocal cords is not critical to the survival of the body, and neither is the function of the thigh muscle. But the function of the thigh muscle is probably more critical, or in other words has a higher degree of criticality, than the function of the vocal cords. This is because, at least in the case of most organisms and most environments, it is harder for an organism to survive if it cannot run or walk than if it cannot utter sound. An element’s function can also be critical for the continued functioning of a system at levels below that of the whole organism. Our definition of critical function can be restated with the overarching system as the context:

F is a critical function for system Y if and only if [23]: (i) some element of Y has F as its function; (ii) the continuing to function of organism Y is causally dependent on the continued performing of F by X.

Thus if a chloride channel in a mucous-producing gland has some kind of genetically inherited defect, it can cause the gland to malfunction and produce an excessively thick fluid. If the gland is in the respiratory epithelium, where its function is to produce a thin slime to moisten the surface of the epithelium, this can cause problems in respiration. Or if the gland is in the pancreas, it can cause the pancreatic fluid to be too viscous to leave the pancreas. In this case, the pancreas malfunctions and causes problems related to nutrition-intake. Thus it is possible for one low-level element to be critical not only for one system but for many.

5.2.1 Degrees of Criticality

A system element has a low degree of criticality if the system can still achieve its function if the element is set out of action. For example, the circulatory system still functions if some particular arterial branch is occluded by a thrombus in such a way that it no longer functions to supply certain regions with blood. For in practice the needed blood flow will be provided via collateral arteries. That means that this particular arterial vessel has a low degree of criticality to the circulatory system. The arterial branches of the aorta, on the other hand, have a high degree of criticality to this system. If they are set out of action this does not mean that the circulatory system will stop functioning, but it will be impaired to a much higher degree than in the case of the absence of a smaller branch. And in some cases branches of an aorta being set out of action can issue in the death of the organism. The criticality of a given function to the survival of a given individual is sometimes relative to the individual and to his environment. The function of the thigh muscle is more critical than that of the vocal chords in a society where people must hunt for food. In this respect, an evaluation of functions according to their criticality is similar to an evaluation of functionings according to their success or failure. Note however that criticality of functions can only be relativized in certain limited cases. The function of the heart is always more critical than the function of the thigh muscle. There are other respects in which an evaluation of functions according to criticality overlaps with an evaluation of functionings according to success or failure. There is not enough room here for a comprehensive account of this overlap. Suffice it to mention a few brief points. One is that an element with a critical function probably has a thinner margin within which its functioning can deteriorate from its prototype without ill effects for the organism as a whole, as compared to an element without a critical function. The liver, for example, must realize its function to remove waste much more successfully than the tonsils must realize their function to guard the oropharynx. Another point at which evaluation of the success of a functioning overlaps with evaluation of the criticality of a function is in certain abnormal circumstances. For example, the lungs and kidneys are both elements of systems responsible for the maintenance of the body’s homeostasis. One function of the kidneys is to maintain ion and water balance, which they realize in part by excreting redundant ions in order to avoid acidosis (the pH level becoming more acidic). If the kidneys are unsuccessful in this performance then the lungs take over: they can maintain the blood pH level, making it more alkaline by means of a more intensive gas exchange. But the lungs can substitute for the kidneys’ function in this way only temporarily. The lungs, then, have one function that becomes critical in the unusual circumstance of kidney failure, but the lungs cannot perform this function successfully enough to contribute to the survival-function of the body.

5.3 Critical Functions and the Spatial-functional Hierarchy

Recall that the spatial-functional hierarchy is organized on the basis of two features of the human body: its complex anatomical structure, and the functions that occur because of the processes that this structure allows for. Elements on lower levels are parts of elements on higher levels, and, correspondingly, their functionings contribute to the functionings of the elements on these higher levels. The circulatory system exists one level down from the body as a whole. As such, its functioning (transporting nutrients, waste material, oxygen, and cells among bodily systems) contributes to the survival of the whole body. The heart is an element of the circulatory system one level down: it is both a part of the circulatory system, and it has a function that contributes to its function. We can now see that a correlation emerges between criticality and spatial-functional level. Elements with functions at higher spatial-functional levels are also more critical. In other words, the fewer systems you have to count upward from a function before you reach the function of the body as a whole, the more critical the function is to the whole organism. For example, the brain exists on a high spatial-functional level: there is only one brain in the whole body, and it has a critical function. Each single neuron, on the other hand, exists on a low spatial-functional level and does not have a critical function because it stands in a relation of redundancy to other neurons. This correlation between criticality and spatial-functional level casts light on the way in which redundancy factors into the spatial-functional hierarchy. Briefly, we can say that the lower the spatio-functional level, the fewer examples we find of criticality and the greater the redundancy of functionings. Thus the mutation of one single cell does not cause cancer in normal conditions (which means: where the immune system is functioning successfully). For this we need the presence of the same mutant gene in a multiplicity cells within a single tissue. It should further be noted that it is functions, not elements, that are critical. It might sound strange to say that it is your aorta’s function that is critical to the functioning of your heart rather than your aorta itself. But consider what happens when your aorta is replaced by a prosthesis: it is then the prosthesis that acquires the function of the aorta in the context of your body. Finally, elements may acquire critical functions in special circumstances. One of your two kidneys has a non-critical function in the body’s normal state, but it becomes critical if the contralateral kidney is damaged or removed and nothing else performs its function. Your kidneys taken together, however, do have a critical function. There are clearly many types of criticality. Is the heart more critical than the stomach because the body will die sooner in virtue of a malfunctioning of the heart? An expansion of this account can break down criticality among its different types. For now, however, since we have explored how the criticality of a function goes hand in hand with its placement on the spatial-functional hierarchy, we have what we need to explain the reasoning behind a division of the body into its major systems.

6. How the Body is Demarcated into Bodily Systems

Note that the demarcation lines among bodily systems are to some degree a matter of fiat; they are boundaries inserted by human beings for the purposes of constructing good (predictively powerful) causal theories [24]. There are physical discontinuities, such as membranes, between many of the elements in bodily systems, but as we have noted these discontinuities only provide partial causal isolation. And many elements have functions for more than one system. We suggest that it is some implicit notion of criticality that has provided the basis for the reasoning underlying current classifications of bodily systems. Medical scientists, in delineating such systems, had to take into account the whole body, since it is in the context of the latter that part-whole functions realize their functionings. Accordingly, all of the top-level elements of the body, its bodily systems, are critical to the body’s continued survival. We can see that the bodily systems are interconnected in such a way that if one system ceases to function then so also, by virtue of the ensuing death of the whole organism, do all the other systems. The death of one is the death of all. However, there is a certain sequentiality to this interdependence, so that the pathologist is in the overwhelming majority of cases able to establish which system was responsible for causing the organism’s life processes to cease. Consider how this applies to the regulatory systems. The autonomous nervous system, which is a regulatory system of the vegetative functions of the body, has as a critical element the brain stem (hypothalamus). A mild stroke in the area of the hypothalamus, where the vegetative centres of regulation are localized, is life-threatening. Another regulatory system is the endocrine system, of which the pancreas is a critical element. If the pancreas does not realize its function to excrete insulin, the organism will not be able to use glucose, and it will die. The immune system has as critical elements T-lymphocytes, whose function is to kill alien cells. If this function is not performed, the organism can die of something as simple as sepsis caused by the saprophytic microflora that normally inhabit the lungs and the intestines (as in the case of AIDS, for example). Our approach suggests also how we might formulate an explanation of the reason why the standard roster of bodily systems, while including certain commonalities, still differ among themselves in certain specific ways. Some textbooks of anatomy, for example, include both bones and joints in the skeletal system, whereas both the Nomina [25] and the Terminologia Anatomica [6] represent bones and muscles as two separate systems. As we saw, there is a certain sequentiality to the interdependence of bodily systems. If one system ceases to function, then others will follow in its train and in a certain order. If two putatively distinct systems always cease to function simultaneously – as in the case of the pulmonary and the systemic components of the circulatory system – then they may for this reason be counted as parts of the same system rather than as systems in their own right. Do the bones and muscles constitute two separate systems or only one? To answer this question is to answer the question whether one can fail without the other thereby failing also. We now wish to assert the hypothesis that all critical functions performed by elements of the body’s hierarchical organization at lower spatial-functional levels are contributions to the performance of critical functions by larger systems on higher levels. Eventually we reach some maximal level, where we are dealing with critical functions belonging to elements that contribute to the functioning of no larger system of the body than the body as a whole. The elements on this maximal level are precisely the body’s major systems. We can then define:

X is a bodily system for organism Y if and only if: (i) X has a critical function for Y; (ii) X is not a part of any other system that has a critical function for Y.

Bodily systems are in other words the largest elements of the human body that have critical functions. Note that (ii) does not exclude isolated elements of X having critical functions in other systems. This is important in accounting for how the failure of one system can cause the failure of other systems. For example, liver failure causes the osmotic pressure in the blood to fail, which causes wide disturbances in the body’s homeostasis. In addition, the liver cannot produce components of the coagulation system, and generalized hemorrhage will occur. Just as some elements belong to a spatial-functional level that is immediately below that of the system of which they form a part – the heart is an example of such an element, since it is not a proper part of any other element of the circulatory system – so bodily systems belong to the spatial-functional level that is immediately below that of the whole organism. The functions performed by the body’s systems are then given in Table 2.

|System |Function |
|Skin |To separate the internal environment from the external medium |
|Musculoskeletal |To move (including: to main the shape of the body and its movement in confrontation with gravity;|
| |to separate sub-environments inside the body; to maintain the internal organs mechanically) |
|Digestive |To digest (to exchange substances: solid substances in-out, liquids in) |
|Respiratory |To breathe (to exchange substances: gas in-out) |
|Circulatory |To supply all the systems of the organism with blood |
|Nervous |To regulate the movement of the body (somatic part) and the vegetative functions of the internal |
| |organs (autonomous part) |
|Endocrine |To regulate metabolism, growth and development and the sexual differentiation of the organism |
|Immune |To preserve the substantial integrity of the organism |
|Urinary |To urinate (to exchange substances: liquid out) |

Table 2. Bodily systems and their functions

Of course it is possible that, if an element several levels below the body as a whole ceases to function, then the life of the the body itself could be brought to an end. Does this undercut a spatial-functional hierarchy? No; rather it forces us to take into account causal processes that relate one spatial-functional level to another and to some environments. The heart is a critical element of the circulatory system; the circulatory system is a critical element of the whole body. If the heart stops, the body dies. But it is not the heart’s stopping that directly causes the body to die; rather, the heart’s stopping causes the circulatory system to stop functioning, which in turn is what causes the body to die. So an element on a lower spatial-functional level, separated from the body as a whole by several other levels, does not directly cause the body to stop working. It does so only by means of intermediate causal links. A spatial-functional hierarchy accounts for these links.

7. Conclusion: Evidence for this Account

The first piece of evidence for the correctness of our account is that it yields a roster of bodily systems that corresponds very well to those given in the standard reference sources. Such sources do not, for example, classify the visual and other perceptual systems as bodily systems alongside those given in our list above. Our analysis enables us to understand also why there is no shared opinion on how to classify the reproductive system within such standard rosters. Some accounts tack the reproductive system onto the urinary system and refer to one composite ‘urogenital system’; some accounts refer to a ‘genital system;’ and some accounts do not mention reproduction at all. We see this as additional evidence for the correctness of our analysis, which should cast light not only on what is broadly shared by standard rosters of the body’s systems but also on the ways in which these rosters differ among themselves. Clearly the reproductive system does not have a critical function in maintaining the body’s life processes (though it might be said that it does have a function critical for the survival of the species, if our account of system turned out to be applicable to systems outside the locus of the body itself). The reproductive system differs further from the other systems in that it comes in two, mutually complementary forms. For its functioning we need individuals of two sexes, each of which contains only part (a half) of the system as a whole. Our approach suggests how we might formulate an explanation of the reason why some textbooks of anatomy include both bones and joints in the skeletal system, while others, including both the Nomina [25] and the Terminologia Anatomica [6], represent bones and joints as two separate systems. We have sought to set out some ontological tools for providing an analysis of ‘bodily system,’ in a way that will do justice to the way the term is used in existing standard sources, while at the same time providing the necessary degree of formal precision to form the basis for a future domain ontology of functional anatomy.

Acknowledgements: This work was supported by the Alexander von Humboldt Foundation under the auspices of its Wolfgang Paul Program. Our thanks go also to Anand Kumar for helpful comments.

References

[1] Smith B. Ontology. In: L. Floridi (ed.), Blackwell guide to philosophy, information and computers, Oxford: Blackwell, 2003.

[2] Baader F. et al. The description logic handbook. Cambridge: Cambridge University Press, 2003.

[3] Borgo S, Gangemi A, Guarino N, Masolo C, Oltramari A. Ontology roadMap: Ontology infrastructure for the Semantic Web, http://wonderweb.semanticweb.org/deliverables/documents/D15.pdf.

[4] Grenon P, Goldberg L, Smith B. Biodynamic Ontology: Applying BFO in the biomedical domain. In: D. Pisanelli, Ontologies in Medicine: Proceedings of the Workshop on Medical Ontologies, Rome, October 2003. Amsterdam: IOS Press, forthcoming.

[5] Structural Informatics Group at the University of Washington Department of Biological Structure; and Division of Biomedical and Health Informatics, Department of Medical Education and Biomedical Informatics. Digital Anatomist Foundational Model, http://sig.biostr.washington.edu/projects/fm/index.html

[6] Terminologia anatomica: international anatomical terminology, Federative Committee on Anatomical Terminology (FCAT), Stuttgart: Thieme, 1998.

[7] World Health Organization training course on National Library of Medicine classification, http://www.emro.who.int/HIS/VHSL/Doc/NLM.pdf.

[8] Köpf-Maier P, ed., Wolf-Heidegger’s atlas of human anatomy, Vol. 1, 5th Edition, Berlin, 2000.

[9] Bertallanfy L. General system theory, New York: George Braziller, 1968. p. 17.

[10] Smith B. Basic formal ontology, http://ontology.buffalo.edu/bfo, 2003.

[11] Ingarden R. Man and value, Munich: Philosophia, 1983, p. 87.

[12] Grenon P, Smith B. SNAP and SPAN: Towards dynamic spatial ontology. In: Spatial cognition and computation, 4:1, forthcoming.

[13] Aristotle. On the parts of the animals. Translated by: Ogle, W. In: http://www.4literature.net/Aristotle/Parts_of_Animals/

[14] Bittner T, Smith B. A theory of granular partitions. In: Foundations of geographic information science, London: Taylor & Francis, 2003.

[15] Rosse C, Mejino JL. A reference ontology for bioinformatics: the Foundational Model of Anatomy. Journal of biomedical Informatics, 2003. In press.

[16] Searle J. The construction of social reality, New York: Free Press, 1995.

[17] Millikan RG. Language, thought, and other biological categories. Cambridge, MA: MIT Press, 1984.

[18] Bigelow J, Pargetter R. Functions. In: The Journal of philosophy 84, 1987, pages 181-196.

[19] Johansson I, Tsikolia N, Ernst D, Elsner K, Siebert D, Munn K, Smith B. Bi-ontological functional anatomy: a new proposal for medical data integration. Under review for the International Workshop on Data Integration in the Life Sciences, 2004, Leipzig, Germany.

[20] Buller J. Etiological theories of function: a geographical survey. In: Biology and philosophy 13, 1998, pages 505-527.

[21] Donnelly M. On holes and parts: The spatial structures of the human body. IFOMIS Reports 03/03, Leipzig, Germany, 2003.

[22] Johansson I. Functions, function concepts, and scales. Forthcoming in: The Monist 1, 2004.

[23] Smith B, Papakin I, Munn K. Bodily systems and the modular structure of the human body. In: Proceedings of AIME 2002: 9th Conference on Artificial Intelligence in Medicine Europe.

[24] Smith B, and Varzi A C. Fiat and bona fide boundaries. In: Philosophy and phenomenological research, 2000, page 60.

[25] Nomina anatomica, 4th ed. Amsterdam: Excerpta Medica, 1977.
-----------------------
[1] See [16], [17], [18]
[2] Philosophers of science who support etiological theories of biological functionality have claimed to have successfully defined (intrinsic) biological function (see, e.g., [21]). There are enough who disagree with this claim, ourselves included, to take it as by no means given that intrinsic functions have been successfully defined.