=Paper=
{{Paper
|id=Vol-2137/paper_16.pdf
|storemode=property
|title=Causation and the River Flow Model of Diseases
|pdfUrl=https://ceur-ws.org/Vol-2137/paper_16.pdf
|volume=Vol-2137
|authors=Fumiaki Toyoshima,Riichiro Mizoguchi,Mitsuru Ikeda
|dblpUrl=https://dblp.org/rec/conf/icbo/ToyoshimaMI17
}}
==Causation and the River Flow Model of Diseases==
https://ceur-ws.org/Vol-2137/paper_16.pdf
Causation and the River Flow Model of Diseases
Fumiaki Toyoshima1*, Riichiro Mizoguchi2* and Mitsuru Ikeda1*
1
Graduate School of Advanced Science and Technology, Japan Advanced Institute of Science and Technology, Nomi, Ishi-
kawa, Japan.
2
Research Center for Service Science, Japan Advanced Institute of Science and Technology, Nomi, Ishikawa, Japan.
ABSTRACT the notion of causation is specified enough, much less
This paper furthers the River Flow Model (RFM) of diseases based on RFM's relationship with OGMS's dispositional account of
the functional view of causation that we have elaborated for the sake of disease. Although it has been long investigated among phi-
formal ontology. The clarification of RFM enables us to offer a clear com-
parison between RFM and the dispositional model of disease given by the losophers, causation still remains an enigma despite its
Ontology for General Medical Science (OGMS) and to see the core ideas enormous importance for ontological modelling in a wide
of disease that RFM and OGMS have in common. This work will be an range of domains. [23]'s argument over causation leaves
initial step towards the development of an ontological module for generic something to be desired, for instance.
disease representation. In this paper we provide further development of RFM on
the basis of the theory of causation that we have recently
1 INTRODUCTION tailored for formal ontology, based on YAMATO: the func-
A disease ontology aims to meet a high demand for a com- tional view of causation [27].2 The elucidation of RFM with
mon framework in which an increasing amount of medical the functional view facilitates the comparison between
information and data are shareable among different infor- OGMS's and RFM's definitions of disease, thereby reveal-
mation systems. A viable definition of disease is thus indis- ing the nature of disease that is conceived in common: dis-
pensable for the robust construction of disease ontologies. ease as a 'causal pattern'. The work done in the paper will be
Designed to represent the entities that are involved in a clin- a first step towards the elaboration of an ontological module
ical encounter in compliance with the framework of the for representing disease.
Open Biomedical Ontologies (OBO) Foundry [25] and The paper is organized as follows. Section 2 presents an
Basic Formal Ontology (BFO) [26], the Ontology for Gen- overview of YAMATO and the functional view of causation.
eral Medical Science (OGMS) [24] offers a general model Section 3 offers a basic idea of RFM, then providing a de-
of disease according to which a disease is a disposition (i) to veloped version of RFM's definition of disease. Section 4
undergo pathological processes that (ii) exists in an organ- compares the RFM of diseases with OGMS's dispositional
ism because of one or more disorders in that organism. In model of disease from the viewpoint of causation. Section 5
marked contrast to a naïve conception of disease, e.g., as a gives concluding remarks and outlines future work.3
state in existing disease ontologies, OGMS's carefully con-
structed definition of disease as a disposition is nowadays 2 THE FUNCTIONAL VIEW OF CAUSATION
utilized in various application ontologies. Causation has been explored in formal ontology directly by
The River Flow Model (RFM) of diseases was initially [7, 12] and indirectly by [2]. There are arguably two prima-
presented in [18] and further developed in [23] as an alter- ry desiderata for a theory of causation that is well suitable
native model of disease that is built in compliance with Yet for formal ontology. One is the generality of causation and it
Another More Advanced Top-level Ontology (YAMATO) requires a philosophically theoretical investigation into cau-
[13]. At the crux of RFM's conception of disease is the idea sation, especially what grounds causation. The other is the
that a disease is a dependent continuant constituted of causal specificity of causation and it demands an ontological
chains of abnormal states. The effectiveness of RFM is veri- framework in which domain experts are able to represent
fied through its high competence in enhancing the interop- various causal phenomena appropriately.
erability and flexibility of disease-related data and infor- We have recently developed the functional view of causa-
mation [11, 28-31].1 tion [27] against an ontological background of YAMATO
A remaining problem with the RFM of diseases resides in (Section 2.1). The functional view aims to meet those two
its explicit reference to the notoriously difficult notion to criteria mentioned above. As for the generality of causation,
model: causation. RFM will not be articulated clearly until the functional view provides the functional grounding of
* 2
To whom correspondence should be addressed: {fumiaki.toyoshima; In this paper we only give a brief summary of the functional view of
mizo; ikeda}@jaist.ac.jp causation owing to spatial limitations. We would be happy to send a draft
1
The OWL file of the RFM ontology is freely available at the website: of [27] to those who wish to know more details about the functional view.
3
http://rfm.hozo.jp/ (Last accessed on July 3, 2017). For a navigation system For the sake of readability, we will hereinafter use the terms 'OGMS-
for the definitions of diseases that is built based on the RFM of diseases disease' and 'RFM-disease' to refer to the notion of disease defined by
(especially on its core notion of causal chains of abnormal states), refer to OGMS and RFM, respectively. The same remark applies to the categories
Disease Compass available at the website: http://lodc.med-ontology.jp/ defined by BFO and YAMATO: we will employ later the terms 'BFO-
(Last accessed on July 3, 2017). process' and 'YAMATO-process', for instance.
1
Toyoshima, F., Mizoguchi, R., and Ikeda, M.
causation: function grounds causation (Section 2.2). As for behavior of the human heart (as a device) is the process of
the specificity of causation, the functional view offers the pressuring blood: the process of change from the low state
functional square of causal relations, according to which to the high state of the pressure (operand) of the blood that
there are four kinds of causal relations: achievement, allow- has travelled through the human heart.
ing, disallowing, and prevention (Section 2.3). Regarding Given the intimate relationship between change and caus-
the comparison between an OGMS-disease and an RFM- al connections, there must be a causal connection behind B
disease, it is important to see that dispositions as inherently as a change in the operand OP.4 Aiming to explain causation
causal properties, if any, are contributors to causation within in terms of an inherently causal entity, the functional view
the framework of the functional view (Section 2.4). of causation consists in saying that B of Od is an inherently
causal process in virtue of its own functional nature whose
combination of how to achieve and what to achieve [10]
2.1 YAMATO: an overview concretizes into the causal relation between what occurs
YAMATO is fundamentally based on [8]'s view of reality: inside Od and S2 in which OP participates (see [27] for de-
continuants and occurrents are existentially interdependent. tails).
Continuants are the entities that persist in time while under- To illustrate this, suppose for instance that a cutting ma-
going various changes (e.g., a person, a river, a statue, and chine cuts a fish into pieces. A cutting machine plays an
an amount of clay). Occurrents are the entities that are tem- agent role as a cutting device in the fish-cutting context and
porally extended (e.g., walking, a walk, and the flowing of a behavior of the cutting device is the change from the state
water). The major subcategory of continuants is objects (e.g., of being one piece to the state of some pieces of the fish that
a person, a river, and a statue, but not an amount of clay). is processed by the cutting machine. The causal connection
Two prominent subcategories of occurrents are processes behind the behavior of the cutting device is the relation be-
(e.g., walking and the flowing of water) and events (e.g., a tween what occurs inside the cutting device (say the cutting
walk). Most importantly, processes are intrinsically 'ongo- motion with a knife) and the state of being some pieces in
ing', whereas events are in nature 'completed'. Among sub- which the fish participates.
categories of occurrents are states: time-indexed qualitative Consider the function of a cutting machine. Analyzed in
occurrents (e.g., being hungry at time t1, sitting at time t2, terms of [10], the function of a cutting machine is to divide
and speeding at time t3). something (what to achieve) and using a sharp knife refers
Concerning relations in YAMATO, an object participates to a specific 'way' of performing the function to divide (how
in a process, an event, or a state. A process is enacted by to achieve). Cutting is in this respect a conceptual mixture
(made possible by) an object that participates in that process. of what to achieve (to divide) and how to achieve (to use a
Any process constitutes a (unique) event. For instance, John knife).
enacts a walking process, which constitutes the walk event. This conception of function justifies the idea that a cut-
Finally, YAMATO exploits the model of roles (a subtype ting behavior of a cutting device is inherently causal primar-
of dependent continuants) [14, 19] the basic tenet of which ily because cutting involves both to divide something to
is understood through the schema: an entity (potential play- some pieces (what to achieve) and to use a sharp object in
er) plays a role as a role-holder in a context. For instance, order to divide something (how to achieve). The state of
Mary plays a student role as a particular student in the being some pieces in which the fish participates and the
school context. The notion of context is too complex to cutting motion with a knife inside the cutting device are
formulate easily, but it can be elucidated through some ex- concretizations of what to achieve and how to achieve of a
amples. For instance, an object can be seen as a context in cutting behavior of a cutting device, respectively.
which other objects play part roles as part and an occurrent Suppose for another example that, when emitted by a nu-
can be considered as a context in which continuants play clear reactor, an electron e1 impacts another electron e2. E1
participant roles as participants, for example. plays an agent role as a device (say Device1) in the context
of a quantum jump of e2 and a behavior of Device1 is the
2.2 The functional grounding of causation process of exciting e2: the change from the lower state to the
higher state of the energy (operand).5 Being of a functional
The functional grounding of causation is based on the de-
nature, the behavior of exciting e2 by Device1 involves both
vice ontology view of reality. The device ontology [15], to bring e2 to the higher energy level (what to achieve) and
whose initial purpose was to analyze technical artifacts,
to use the motion of e1 (how to achieve).
refers to a specific way of assigning roles to objects in gen-
eral (whether natural or artificial). A device in the device
ontology refers to a role-holder that processes something (i.e. 4
Consider for instance that an object travels at a constant velocity. A be-
receives something to produce something). The device on- havior of the object in constant motion is the change in the position (oper-
tology enables us to analyze as devices a full range of ob- and) corresponding to the transition between two places which the particle
jects (including the human heart and an electron). occupies. The causal connection behind this behavior is the relation be-
According to the device ontology, an object O plays an tween the motion of the object and the position change behavior of the
object.
agent role as a device Od in a given context and a behavior 5
An electron that receives the amount of energy exceeding the threshold
B of Od is the change from one state S1 to the other state S2
value is said to make a quantum jump or to be excited from the lower-
of the operand OP that is processed by Od. For instance, a energy state to the higher-energy state.
2
Causation and the River Flow Model of Diseases
It is interesting to note that, given the device ontology, 2.3.4 Disallowing
the example of a quantum jump of an electron bears a close Disallowing consists in 'preventing' some sufficient condi-
similarity to the scenario in which, when shot by a billiard tion for a phenomenon, thereby preventing the phenomenon.
cue, a billiard ball collides with another billiard ball. As this It is therefore indirect with respect to prevention and it is
shows, the device ontology view of reality is versatile also negative, just as prevention is. Example: ras inactiva-
enough to be applicable to different granularity levels of tion disallows signal amplification of MAP kinease cascade
reality. (by 'preventing' the activation of kinease molecules such as
MAPKK in the following MAP kinease cascades).
2.3 The functional square of causal relations
The functional square of causal relations is the fourfold dis- 2.4 Dispositions as contributors to causation
tinction of causal relations that are in nature functional: Dispositions are of high utility in overall ontological model-
achievement, allowing, disallowing, and prevention (see ling, especially in the biomedical domain [22]. For the sake
Table 1). The functional square claims to be an all- of future argument, we will hereafter use BFO's notion of
encompassing table of causal relations, thereby meshing disposition, which relies on a BFO-process (but not a
well with a counterfactual theory of causation [6, 21]. YAMATO-process): an occurrent that has temporal parts
and always depends on some material entity (an independ-
Table 1 The functional square of causal relations ent continuant that has some portion of matter as part).
A disposition is a dependent continuant (bearer) that ex-
Indirect Direct ists because certain features of the physical make-up (mate-
rial basis) of the independent continuant in which it inheres
Positive Allowing Achievement and whose instances can be realized in associated BFO-
Negative Disallowing Prevention processes of specific correlated types in which the bearer
participates. For instance, fragility is the disposition of a
glass (bearer) to break (realization) that depends on a partic-
2.3.1 Achievement ular physical molecule structure (material basis) of the glass.
Achievement is a direct and positive causal relation: a rela- In philosophy, there is a growing interest in a disposition-
tion between occurrents (not only events but also processes al theory of causation according to which, roughly, causa-
and states) that we ordinarily take to be causal. The func- tion occurs when a disposition realizes itself [20]. The func-
tional view of causation [27] deduces achievement from an tional view of causation says that a realization of a disposi-
inherently causal behavior of a device (Section 2.2) and a tion is a contributor to causation. Recall the example of a
recent work on a unifying definition for artifact and biologi- cutting machine. A cutting device may possess the disposi-
cal functions [3, 16, 17]. tion of a knife whose realization is the cutting motion as a
Example: regarding thrombosis, the growing process of a concretization of how to achieve of cutting. In this way, a
blood clot in a blood vessel achieves the state of being small realization of a disposition serves as a basic ingredient of an
of the cross-sectional area of the blood vessel, which in turn inherently causal behavior of a device. Functional causal
achieves the state of being in short supply of oxygen. Note power is nonetheless irreducible to dispositional causal
that thrombosis is cured only when the cross-sectional area power, as the functional view says.
of the blood vessel is of a clinically normal size.
3 THE RIVER FLOW MODEL OF DISEASES
2.3.2 Allowing
Allowing holds between an occurrent and the state of being 3.1 A basic idea
a necessary condition for achievement. In this respect, al- We said in Section 2.1 that a river is an object (continuant)
lowing is indirect with respect to achievement. It is also and the flowing of water is a process (occurrent). According
positive because its holds both between actual occurrents to YAMATO or rather [8], an object is a unity that enacts its
both ontologically and linguistically. Example: ATP pro- external process or the 'interface' between its internal pro-
duction allows active calcium transport (by achieving the cess and external process. The basic tenet of RFM is the
state of being in operation of transporters as a necessary analogy between a river and a disease. Just as a river enacts
condition for active calcium transport). changing the course of the flow of water as its external pro-
cess, a disease enacts as its external process a process of,
2.3.3 Prevention e.g., spreading and disappearing. While a river is an inde-
Prevention is direct because it is ontologically the same as pendent continuant, however, a disease is a dependent con-
achievement but it is nonetheless negative because, linguis- tinuant: it depends on an organism as its bearer.
tically speaking, it has as relatum a non-actual occurrent Moreover, just as a river has the flowing of water as its
(see [27] for details). Example: a tendon prevents the sepa- internal process (a process that occurs inside the river), a
ration between muscle and bone. Note that this is essentially disease has as its internal process a number of chains of
the same as the achievement by a tendon of the connection causal phenomena. A disease is in this respect constituted of
between muscle and bone. causal chains of phenomena that are harmful to the organ-
ism from a medical viewpoint. This constituted-of relation
3
Toyoshima, F., Mizoguchi, R., and Ikeda, M.
has a close affinity with the constitution relation of a pro- vention, either of which is initiated by at least one abnormal
cess to the event (see Section 2.1). state.
All these considerations lead to the above-mentioned
main idea of the RFM of diseases: a disease is a dependent Note also that an RFM-disease has at least one clinical im-
continuant constituted of causal chains of abnormal states. balance state (see Section 3.2).
Type I diabetes, for instance, may have inside it the causal The reader should keep in mind that RFM's notion of
relation between the state of being broken of pancreatic beta clinical abnormality is virtually primitive. This is justified
cells and the state of being little of insulin in the blood. through the sharp distinction between the domain-neutral
notion of clinical abnormality and the domain-specific task
of the identification of clinical abnormality. What counts as
3.2 The imbalance model clinically abnormal would vary from clinical to clinician in
The imbalance model, initially presented in [18], is a key a broad biomedical field. RFM purports to be an ontological
element of RFM in the sense of supplementing and general- (domain-neutral) model of diseases and RFM's notion of
izing a simple conception of a disease as causal chains of clinical abnormality refers to the existence (rather than the
abnormal states. According to the imbalance model, a clini- content) of criteria for observing a state from a clinical per-
cally normal organism maintains homeostasis (by which we spective. For an application of RFM's notion of abnormal
basically mean a disposition of an organism to regulate its states, see [11, 28-31].
body in close with [24]) when the 'supply’ and the 'demand'
are well-balanced as regards all the parameters that are rele- 4 DISCUSSION
vant to the organism's living condition. In a clinically ab-
normal organism, however, the supply and the demand for 4.1 OGMS's dispositional model of disease
some (if not every) parameter that relates to the organism's To investigate the commonalities and differences between
life are so different from each other that the difference be- RFM and OGMS's dispositional model of disease, we brief-
tween them lies outside the clinically permissible range for ly present some core terms of OGMS. A disorder basically
the maintenance of the organism's homeostasis. refers to a material entity which is clinically abnormal and
To illustrate the imbalance model, consider diabetes in part of an organism, although the precise definition of a
general. In a patient with diabetes (whether Type I diabetes disorder has been repeatedly changed and seems to be under
or steroid diabetes), the required amount of insulin (de- development (see [4, 5, 24]). A pathological process is a
mand) exceeds to a clinically abnormal degree the amount bodily process that is a manifestation of a disorder, where a
available for working insulin (supply) and this imbalance bodily process is a BFO-process in which participate one or
state causes the state of being at an elevated level of glucose more material entities within or on the surface of an organ-
in the blood, which may result in the loss of sight of the ism. Pathological process are recognized through symptoms
patient over a long period of time. The difference between and signs.
Type I diabetes and steroid diabetes resides partly in the fact An OGMS-disease is a disposition (i) to undergo patho-
that, in the case of Type I diabetes and steroid diabetes, logical processes that (ii) exists in an organism because of
causal chains that lead to the imbalance state described one or more disorders in that organism. The material basis
above include the state of having depleted pancreatic beta of a disease as a disposition is a disorder of the disease and a
cells and the state of having large quantity of steroids, re- disease comes into existence when its corresponding disor-
spectively. der does, i.e. when the organism disposes towards its rele-
We have the following definition of a clinical imbalance vant pathological processes. A disease as a disposition may
state: go unrealized, e.g., when it lies dormant over a long period
of time. A disease course is the totality of all BFO-
Clinical imbalance state =def. a state of an organism such processes through which a given disease instance is realized.
that, given a parameter p that is relevant to the organism, the A disease course of a disease ranges widely from potentially
mismatch between the supply and the demand that are speci- asymptomatic early stages of the disease to its recognizable,
fied with respect to p falls outside a clinically normal range pathological processes.
for the organism's homeostasis. For instance, epilepsy as a disease is a disposition to un-
dergo the occurrence of seizures (pathological processes)
that exists owing to some clinically abnormal, neuronal cir-
3.3 The RFM definition of disease
cuitry of the brain (disorder).
An existing RFM definition of disease [18, 23] involves the
term ‘causal chain’. Based on the functional square of caus-
al relations presented in Section 2.3, we have the following 4.2 Disease as a 'causal pattern'
revised and developed RFM definition of disease: RFM and OGMS share some common views on disease.
First of all, an RFM-disease and an OGMS-disease both say
Disease (revised) =def. a dependent continuant that is consti- that a disease is in nature a 'causal pattern'. An RFM-disease
tuted of abnormal states occurring in an organism that are is characterized by the regular way in which abnormal states
connected by achievement, allowing, disallowing, or pre- are causally connected and the pattern nature of an RFM-
4
Causation and the River Flow Model of Diseases
disease may be represented in terms of a directed graph in the same diabetes category because they are essentially the
graph theory [9]. 'diabetes disposition'. These two diseases are nevertheless
Similarly, an OGMS-disease reasonably qualifies as a different because a predisposition to have diabetes that is
causal pattern. For one thing, an OGMS-disease is causal, involved in Type I diabetes is of a different nature from that
since a disposition is an inherently causal property. For an- involved in steroid diabetes. A realization of the former pre-
other, an OGMS-disease is of a pattern nature because a disposition, but not of the latter predisposition, may have as
disposition which has a 'specific' material basis realizes its part the BFO-process of destruction of pancreatic beta cells;
'corresponding' BFO-processes when exercised under some conversely, a realization of the latter predisposition, but not
'appropriate' circumstances. of the former predisposition, may have as part the BFO-
It is important to emphasize that a disease is a dependent process of the increase of steroids.
continuant, but not an occurrent, although some existing
ontologies classify a disease a subtype of occurrent. A dis-
ease is some entity with which a patient is affected and
4.4 Causation: dispositional vs. functional
which medical practitioners identify, diagnose, and cure. A Though conceived as a causal pattern in common, an
disease is something that comes into existence, grows, and OGMS-disease and an RFM-disease are significantly differ-
finally disappears in the patient's body. All these observa- ence in the sense that OGMS's basic ''unit of thought'' is a
tions imply that a disease is an entity that persists in time, i.e. disposition (dependent continuant) but RFM's is a state (oc-
a continuant. Additionally, a disease is a dependent continu- current). This fundamental difference between an OGMS-
ant that inheres in an organism. disease and an RFM-disease is largely, if not totally, expli-
cable in terms of causation.
For OGMS, causation is dispositional: causation occurs
4.3 Clinical threshold when a disposition as an inherently causal property realizes
Furthermore, both an RFM-disease and an OGMS-disease itself. The dispositional conception of causation leads di-
involve what we may call a 'clinical threshold': the level at rectly to an OGMS-disease as a disposition. For RFM, cau-
which symptoms of a disease begin to develop. RFM char- sation is functional: causation occurs when an inherently
acterizes a clinical threshold employing the imbalance mod- causal behavior of an object (a device) achieves some occur-
el. Symptoms of a disease show themselves when the clini- rent. An RFM-disease inherits its causal nature from the
cal imbalance state that is relevant to the disease has oc- linkages, brought about by functional causal power, among
curred. Likewise, OGMS explicates a clinical threshold in abnormal states.
terms of the emergence of a disease as a disposition (and of
its corresponding disorder as a material basis of the disposi-
tion). Having reached a clinical threshold, an organism dis-
4.5 Clinically abnormal: continuant vs. occurrent
poses towards pathological processes, which are recogniza- Another indicator of the contrast between OGMS's and
ble through symptoms. RFM's conceptions of causation is the difference in what is
It is nonetheless vital to clarify the difference between an clinically abnormal between OGMS and RFM. Clinically
RFM-disease and an OGMS-disease from the viewpoint of a abnormal are a disorder (continuant) in OGMS and a state
clinical threshold. In the case of an RFM-disease, a clinical (occurrent) in RFM. In OGMS, a disease as a disposition
imbalance state of a disease is not always an initial state of inherits its clinical abnormality from a disorder as its mate-
the disease. As said in Section 3.2, the imbalance model rial basis; and therefore, a disorder (a material entity) is de-
abstracts from a disease its generality and eliminates its fined as clinically abnormal. In contrast, RFM's notion of
specificity. For instance, Type I diabetes and steroid diabe- clinically abnormal state reflects well the idea of an inher-
tes fall into a group of diabetes, since they have the same ently causal occurrent (process) embraced by the functional
kind of clinical imbalance state (i.e. the deficiency of insu- view of causation.
lin), but those two diseases still differ from each other be- Consider for instance inflammation as a disease. OGMS
cause they have different causal chains of abnormal states. would say that inflammation as a disposition is clinically
By comparison, it is clear that the emergence of an abnormal primarily because so is its material basis, e.g., the
OGMS-disease as a disposition is always at the beginning of cells in the relevant part of the organism. RFM says howev-
the disease. To do justice to the specificity as well as the er that the clinical abnormality of the cells there, if any, is a
generality of disease, OGMS covers a predisposition to dis- contributor to that of states that inflammation has inside it.
ease of type X: a disposition in an organism that constitutes This marks a close analogy with the argument for the func-
an increased risk of the organism's subsequently developing tional view of causation: a realization of a disposition con-
the disease X [5, 24]. A predisposition is a disposition to tributes to functional causation (see Section 2.4).
acquire a further disposition and some diseases as disposi-
tions (e.g., osteoporosis) are predispositions to further dis- 5 CONCLUDING REMARKS
eases as dispositions (e.g., fracture). We have furthered the River Flow Model (RFM) of diseases
Roughly speaking, the generality and the specificity of an employing the functional view of causation. By so doing,
OGMS-disease are to be captured by a disease as a disposi- we have offered a clear comparison between an RFM-
tion and a predisposition to disease of type X, respectively. disease and an OGMS-disease in terms of causation, in par-
For instance, Type I diabetes and steroid diabetes belong to ticular the contrast between functional causation and dispo-
5
Toyoshima, F., Mizoguchi, R., and Ikeda, M.
sitional causation. The work done here will contribute to the Conference of Formal Ontology in Information Systems (FOIS 2004)
formalization of a general disease module for foundational (pp. 273-284). IOS Press.
ontologies (whether BFO, YAMATO, or others) and also to 13. R. Mizoguchi (2010). YAMATO: Yet Another More Advanced Top-
the methodological supplementation of OGMS [1]. level Ontology. In Proceedings of the Sixth Australasian Ontology
Future work includes further development of RFM using Workshop (pp. 1-16). University of Adelaide.
the functional view of causation. For instance, the explora- 14. R. Mizoguchi, A. Galton, Y. Kitamura, and K. Kozaki. (2015). Fami-
tion of the formal relationships among the causal relations in lies of roles: A new theory of occurrent-dependent roles. Applied On-
the functional square would enable us to have a closer ex- tology 10 (3-4): 367-399.
amination of the nature of causal chains of abnormal states 15. R. Mizoguchi and Y. Kitamura (2009). A Functional Ontology of
in a disease. Along another line of research lies the question Artifacts. The Monist 92 (3): 387-402.
of whether RFM can be extended to mental disease, just as 16. R. Mizoguchi, Y. Kitamura, and S. Borgo. (2012). Towards a Unified
OGMS's disease model was in [4]. To address this question Definition of Function. In Proceedings of the 7th International Con-
would demand the task of investigating mental causation, ference of Formal Ontology in Information Systems (FOIS 2012) (pp.
which is currently outside the scope of the functional view. 103-116). IOS Press.
17. R. Mizoguchi, Y. Kitamura, and S. Borgo. (2016). A unifying defini-
ACKNOWLEDGEMENTS tion for artifact and biological functions. Applied Ontology 11 (2):
We are deeply grateful to Yuki Yamagata for her valuable 129-154.
comments on a variety of causal phenomena in biomedicine. 18. R. Mizoguchi, K. Kozaki, H. Kou, Y. Yamagata, T. Imai, K. Waki,
and K. Ohe. (2011). River Flow Model of Diseases. In Proceedings of
REFERENCES the 2nd International Conference on Biomedical Ontology
(ICBO2011) (pp. 63-70).
1. A. Barton, A. Rosire, A. Burgun, and J. F. Ethier. The Cardiovascular
19. R. Mizoguchi, E. Sunagawa, K. Kozaki, and Y. Kitamura. (2007).
Disease Ontology. In Proceedings of the 8th International Confer-
The model of roles within an ontology development tool: Hozo. Ap-
ence of Formal Ontology in Information Systems (FOIS 2014) (pp.
plied Ontology 2 (2): 159-179.
409-414). IOS Press.
20. S. Mumford and R. L. Anjum (2011). Getting Causes From Powers.
2. A. Barton, R. Rovetto, and R. Mizoguchi. (2014). Newtonian Forces
Oxford University Press.
and Causation: A Dispositional Account. In Proceedings of the 8th
21. L. A. Paul and N. Hall. (2013). Causation: A User's Guide. Oxford
International Conference of Formal Ontology in Information Systems
University Press.
(FOIS 2014) (pp. 157-170). IOS Press.
22. J. Röhl and L. Jansen. (2011). Representing dispositions. Journal of
3. S. Borgo, R. Mizoguchi, and Y. Kitamura (2016). Formalizing and
Biomedical Sementics 2 (Suppl 4): S4.
Adapting a General Function Module for Foundational Ontologies. In
23. R. J. Rovetto and R. Mizoguchi. (2015). Causality and the ontology
Proceedings of the 9th International Conference of Formal Ontology
of disease. Applied Ontology 10 (2): 79-105.
in Information Systems (FOIS 2016) (pp. 241-254). IOS Press.
24. R. H. Scheuermann, W. Ceuster, and B. Smith. (2009). Towards an
4. W. Ceusters and B. Smith (2010). Foundations for a realist ontology
Ontological Treatment of Disease and Diagnosis. In Proceedings of
of mental disease. Journal of Biomedical Semantics 1 (10).
the 2009 AMIA Summit on Translational Bioinformatics (pp. 116-
5. W. Ceusters and B. Smith (2010). Malaria Diagnosis and the Plasmo-
120). San Francisco, CA.
dium Life Cycle: The BFO Perspective. In Interdisciplinary Ontology.
25. B. Smith et al. (2007). The OBO Foundry: coordinated evolution of
Proceedings of the Third Interdisciplinary Ontology Meeting (pp. 25-
ontologies to support biomedical data integration. Nature Biotechnol-
24). Keio University Press.
ogy 25: 1251-1255.
6. J. Collins, N. Hall, and L. A. Paul. (eds.) (2004). Causation and
26. B. Smith et al. (2015). Basic Formal Ontology (BFO): Specification
Counterfactuals. MIT Press.
and user's guide. Available online at: https://github.com/bfo-
7. A. Galton (2012). States, Processes and Events, and the Ontology of
ontology/BFO/wiki (Last accessed on July 3, 2017).
Causal Relations. In Proceedings of the 7th International Conference
27. F. Toyoshima and R. Mizoguchi. (under review). Causation: its func-
of Formal Ontology in Information Systems (FOIS 2012) (pp. 279-
tional grounding. Applied Ontology.
292). IOS Press.
28. Y. Yamagata, T. Imai, K. Ohe, and R. Mizoguchi. (2014). Towards
8. A. Galton and R. Mizoguchi (2009). The Water Falls but the Water-
the Integration of Abnormality in Diseases. In Proceedings of the 5th
fall does not Fall: New perspectives on Objects, Processes and Events.
International Conference on Biomedical Ontology (ICBO2014).
Applied Ontology 4 (2): 71-107.
29. Y. Yamagata, H. Kou, K. Kozaki, R. Mizoguchi, T. Imai, and K. Ohe.
9. J. M. Harris, J. L. Hirst, and M. J. Mossinghoff. (2000). Combinator-
(2013). Ontological Modeling of Interoperable Abnormal States. In
ics and Graph Theory. Springer.
Proceedings of Second Joint International Semantic Technology Con-
10. Y. Kitamura, T. Sano, K. Namba, and R. Mizoguchi. (2002). A Func-
ference (JIST 2012), vol. 7774 (pp. 33-48). Springer.
tional Concept Ontology and Its Application to Automatic Identifica-
30. Y. Yamagata, H. Kou, K. Kozaki, R. Mizoguchi, T. Imai, and K. Ohe.
tion of Functional Structures. Advanced Engineering Informatics 16
(2013). An Ontological Model of Abnormal States and its Application
(2): 145-163.
in the Medical Domain. In Proceedings of the 4th International Con-
11. K. Kozaki, R. Mizoguchi, T. Imai, and K. Ohe. (2012). Identity
ference on Biomedical Ontology (ICBO 2013).
Tracking of a Disease as a Causal Chain. In Proceedings of the 3rd
31. Y. Yamagata, K. Kozaki, T. Imai, K. Ohe, and R. Mizoguchi. (2014).
International Conference on Biomedical Ontology (ICBO 2012).
An ontological modeling approach for abnormal states and its appli-
12. J. Lehmann, S. Borgo, C. Masolo, and A. Gangemi. (2004). Causality
cation in the medical domain. Journal of Biomedical Semantics 5: 23.
and Causation in DOLCE. In Proceedings of the 3th International
6