-
Causation is of cardinal importance to the biomedical field. For
one thing, causation and causal inference play a central role in
sciences in general [
1
], let alone biomedicine [
2,3
], because
scientists typically observe the natural world to search for causes
of phenomena therein and provide a causal explanation of them.
For another, closer examination of the causal relationships
between pathological processes helps to provide a solid
foundation for a generic ontological model of disease [
4,5
] and
to enhance an efficient computational extraction of biological
data and information [
6
].
Despite some prior works [
4,5
], it is an open question how
biomedical specialists should conceptualize causation and
represent it formally. Rather, causation still remains an enigma
to us, although it has been carefully investigated in a number of
different domains, ranging from philosophy [
7,8
] and
foundational ontology research [
9-12
] to artificial intelligence
[
13
], linguistics [
14
], and cognitive science [
15
]. It is
nonetheless a pressing issue to build an ontology of causation to
serve as a common semantic framework which would facilitate
the annotation of biomedical datasets.
In this paper we endeavor to carry out an extensive ontological
exploration of causation vis-à-vis its practical usage in the
context of biomedical ontologies. Granted that philosophical
ontology proves to be beneficial to a robust construction of
scientific ontologies, including biomedical ones [
16,17
],
causation belongs to the notions that are loosely grouped under
the heading of ‘natural necessity’: e.g., necessarily, a glass
breaks when it is pressed with a certain force. Other relevant
topics of natural necessity include laws of nature, dispositions,
and counterfactuals (see [
18
] for details).
In this general setting, a dispositional approach to causation is
arguably most suitable to meet the needs of those who engage
actively in biomedical ontologies. Based on this stipulation, our
inquiry begins by considering a dispositional interpretation of
what we call ‘canonical causation’: arguably the most typical
kind of causal relation that holds between particular processes.
We then extend this dispositional understanding to other kinds
of causation that are commonplace in the biomedical practice as
well as in other situations. We also discuss a dispositional
perspective on counterfactuals, namely another notion of natural
necessity than causation, and also on scientifically relevant
epistemic notions such as (scientific) explanations and
hypotheses. Our work will contribute to offering a vivid
ontological picture of the biomedical world.
For the sake of the anchoring of a general ontological
background, we deploy Basic Formal Ontology (BFO) [
19,20
]
as an upper ontology (aka foundational ontology), namely an
ontology to furnish the most general categories (e.g., space and
time) and relations (e.g., identity and parthood) to serve as a
useful guideline for building domain ontologies of high semantic
interoperability. The BFO-based investigation of causation
would be valuable for biomedical ontologies because the
practical utility of BFO to them is shown by a widespread
recognition of the relevance of the BFO methodological
principle of ontological realism [
21
] (which says that the most
effective way to ensure well-founded scientific ontologies is to
view ontologies as representations of the reality that is described
by science) to development of biomedical ontologies [
22
]; and
by the achievement of the Open Biomedical Ontologies (OBO)
Foundry [
23
], i.e., a collaborative project to coordinate
ontologies to support biomedical data integration such that BFO
can provide a common semantic basis for all the OBO
ontologies.
In the most basic BFO framework, entities fall into two kinds:
universals (aka types, classes) and particulars (aka tokens,
instances). Particulars (e.g., Mary) bear the instance-of relation
to universals (e.g., Human). Particulars, on which we put a
central focus in this paper, fall into two categories: continuants
and occurrents. Characteristically, continuants can persist, that is
Copyright © 2019 for this paper by its author. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
to say, they can exist at one time and also exist at another
different time; whereas occurrents (including processes) extend
through time. Continuants can be further divided into
independent continuants (including objects) and dependent
continuants (intuitively: properties). Independent continuants, or
especially objects (e.g., stones) can be bearers of dependent
continuants (e.g., hardness) and they can also participate in
occurrents (e.g., a fall of the stone).
This paper is organized as follows. Section II argues for a
dispositional ontology of causation that would be maximized in
the biomedical field and provides a dispositional construal of
canonical causation and quasi-causation in our terminology.
Section III considers dispositionally other varieties of causation:
probabilistic causation, simultaneous causation, and mental
causation. Section IV discusses a dispositional view of
counterfactuals, explanations, and hypotheses. Section V
concludes with some brief remarks on future work.
II. The Dispositional Grounding of Causation
There is broad agreement among prior works [
9-12
] in
foundational ontology research on some basic characteristics of
causation, in spite of its complicated character. First, causation
is a binary relation between processes. The terms ‘causation’
and ‘causal relation’ will be hereafter used interchangeably.
When pressing with a certain force caused a glass to be broken,
for instance, the process of the pressing the glass with a certain
force has the causal relation to the process of the breaking of the
glass. Second, there are two types of causation: type-level
causation (e.g., smoking causes lung cancer) and token-level
causation (e.g., Mary’s smoking caused her to get lung cancer).
Lehmann et al. [
9
] call the former and the latter ‘causality’ and
‘causation’, respectively; and causality and causation refer to the
causal relation between process universals and between process
particulars, respectively. (The terms ‘type-level causation’ and
‘token-level causation’ are preferred to be employed in this
paper to circumvent terminological confusion, though.) In this
paper we focus mainly on token-level causation. Third,
causation has a significant bearing on scientifically important
epistemic notions such as explanation and hypothesis (see
Section IV for details).
We begin by focusing on the most paradigmatic kind of
causation (which we call ‘canonical causation’) and then
consider other variants of causation in Section III. Although it
may defy easy description, canonical causation possesses at least
the following four features. First, it is token-level causation that
is observed in the macroscopic world in which classical
(Newtonian) mechanics holds. Throughout this paper, we
postulate this worldview to avoid complications added by a
contentious ontological role of causation in non-classical
physical theories such as general relativity and quantum
mechanics. Second, it is so-called forward causation (where the
cause occurs earlier than its effect); and backwards causation
[
24,25
] (where the cause occurs later than its effect) is outside
the scope of our investigation. Third, it is so-called physical
causation (which is, roughly, entirely explicable purely in
physical terms). Fourth and lastly, it is non-probabilistic. For
instance, it is always the case that pressing with a certain force
caused a glass to be broken, and in this sense, the process of
pressing a glass with a certain force bears the canonical causal
relation towards the process of the glass being broken. In
contrast, when a toss of a coin caused the coin to land on heads,
the causal relation between the process of the toss of the coin
and the process of the coin landing on heads is probabilistic
because the coin could have landed on tails (with the probability
50 percent) [
26
].
As for the formal properties of the (canonically) causal relation,
it is irreflexive and asymmetric. That is to say, no process
caused itself, and if a process x caused a process y, then it was
not the case that y caused x. On the other hand, the transitivity of
the causes relation (i.e., whether a process x caused a process z if
x caused a process y, which in turn caused z) is not presupposed
because it is a highly debatable subject [8, Chapter 5] [
27
].
It has been pointed out in the relevant literature [
11
] that there is
a causal-like relation which a state bears to serve as a
precondition for the causal relation to hold that and which we
may call ‘quasi-causation’. For instance, the state of oxygen
bears the quasi-causal relation towards the causal relation
between striking of a match and the lighting of the match. This
quasi-causal relation can be labeled as the verb ‘allow’ [
11
]:
e.g., the state of oxygen allowed the striking of a match to cause
the match to light.
Concerning states, Galton [
28
] identifies their two distinct, albeit
related, meanings:
•
•
States as continuants. An ‘instantaneous state’ of some thing or situation, as given by the values assumed at one time by some of its variable properties. E.g., the position and momentum of a particle in physics.
States as occurrents. A ‘state situation’, described as unchanging with respect to some selected. property or combination of properties. E.g., the state of the water temperature being 50 degrees Celcius.
Note that this general analysis of states is useful in biomedicine,
as is indicated by its concordance with Chaudhri and Inclezan’s
[
29
] observation of the multiple usages of the term ‘state’ in a
biology textbook.
B)
Theory of Dispositions
A disposition is a (specifically) dependent continuant that is
linked to a realization, namely to a specific possible behavior of
an independent continuant that is the bearer of the disposition.
To be realized in a process, a disposition needs to be triggered
by some other process. Classical examples include fragility (the
disposition to break when pressed with a certain force) and
solubility (the disposition to dissolve when put in a certain
solvent). Characteristically, dispositions may exist even if they
are not realized or even triggered. A glass is fragile even if it
never breaks or even if it never undergoes any shock, for
instance. We will also speak of a categorical basis [
30
] (aka a
base [
31
]) and a background condition [
31
] of a disposition: a
quality (or a sum of qualities) of the disposition bearer and a
necessary condition for the realization of the disposition,
respectively. For instance, flammability of this match is the
disposition to be realized when the match is struck (trigger)
against a certain surface in an oxygenated environment
(background condition), thereby bringing about the production
(realization) of fire; and it is based on a particular molecule
structure (categorical basis) of the match.
There are nowadays several accounts of the identity of
dispositions available. For instance, Barton, Grenier, Jansen and
Ethier [
30
] criticize Röhl and Jansen’s [
31
] traditional
disposition model for being susceptible to ‘disposition
multiplicativism’: the excessive and arbitrary proliferation of
dispositions that would nullify their ontological importance.
They instead propose an alternative model of dispositions: two
dispositions are identical if and only if (‘iff’ hereafter) they have
the same categorical basis, the same universal of minimal
triggers, and the same universal of maximal realizations, where
the universal of minimal triggers of a disposition d is the
universal of triggers of d for which no proper part is a trigger of
d, and the universal of maximal realizations of d is the universal
of realizations of d which are not proper parts of another
realization of d. For instance, a minimal trigger and a maximal
realization of the fragility disposition of a glass would be the
weakest shock on the glass that is strong enough to enable the
glass to break and the whole process of the breaking of the glass,
respectively. To keep things manageable, we will employ a
practical identity condition [
32
] of dispositions that is acceptable
in both the above-introduced frameworks for dispositions: two
dispositions are identical iff they have the same categorical
basis, the same universal of triggers and the same universal of
realizations.
We contend that a dispositional understanding of causation can
be most fully exploited in the biomedical domain as compared to
e.g., its primitivism [
33
] and its lawful interpretation [
34
]. For a
negative reason, even the existence of biological laws is a highly
controversial topic [
35
]. For a positive reason, realizable entities
such as tendencies and dispositions are central to medical
information sciences [
36
]; and dispositions serve as such a
useful conceptual tool for the analysis of the explanatory
practice in the biological sciences [
37
] that a dispositional theory
of causation captures well the dynamicity, continuity, and
context-sensitivity of biological phenomena [
38
]. It has been
also argued that a dispositional analysis of causation helps to
contribute to evidence-based medical practice [
39
] than its
counterfactual analysis [
40,41
] (but see [
42
] for criticism).
Moreover, a dispositional approach to causation has its
advantages with regard to OBO ontologies. For one thing, BFO
is fundamentally committed to dispositions, as is indicated by its
explicit grounding of natural necessity in dispositions:
“Incorporation of dispositions into the BFO ontology provides a
means to deal with those aspects of reality that involve
possibility or potentiality without the need for complicated
appeals to modal logics or possible worlds” [20, p. 102]. For
another, dispositions have been carefully investigated in
ontology research in general ontology research for the last
decade. Accordingly, they have been conceptually and logically
examined, so that ontology of dispositions has been exploited
for formalizing various biomedical entities: e.g., diseases
[
32,43,44
], medical risk [
45
], and medical Bayesian indicators
of performance [
46
].
The crux of a dispositional approach to causation is that when
causation occurs, there is a corresponding realization of some
disposition [
47
]. A glass broke, as the dispositional theory goes,
in virtue of the realization of the fragility disposition of the glass
triggered by pressing the glass with a certain force. The causal
relation between two processes can be thus interpreted as the
relation between the triggering and the realization processes of
some disposition.
Furthermore, quasi-causation would be explicable in terms of a
background condition of a disposition. For one thing, the trigger
and a background condition of a certain disposition are largely
pragmatically distinguished causal factors [
31
] and this coheres
well with the idea that quasi-causation is so relevant to ontology
of causation that a certain causal relation would cease to hold if
any one of preconditionsl states for the causal relation failed to.
For another, background conditions can comprise categorially
diverse kinds of entities [
31
] and this conforms to the
aforementioned observation that preconditional states of
quasicausation can be construed as either continuants or occurrents.
The state that brings about quasi-causation can be therefore
explicated in terms of a background condition of some
disposition.
Against the BFO background, a ‘continuant-state’ can be
construed as a subtype of quality (see e.g., [
48
]): “A specifically
dependent continuant that, if it inheres in an entity at all, is fully
exhibited, manifested, or realized in that entity. In order for a
quality to exist, one or more independent continuants must also
exist. Examples include the mass of a kidney, the color of this
portion of blood, and the shape of a hand” [20, p. 183]. In
contrast, an ‘occurrent-state’ would be as a subtype of process:
“An occurrent entity that exists in time by occurring or
happening, has temporal parts, and always depends on at least
one independent continuant as participant” [ibid.]. Since it
essentially needs some participant, an occurrent-state (process)
of a participant occurs in virtue of the fact that the participant
bears some relevant continuant-state (quality).
III. Other Varities of Causation
A)
Probabilistic Causation
The exemplary dispositions (e.g., fragility and flammability) that
we have so far discussed are all ‘sure-fire dispositions’: “such
dispositions that will necessarily be realized given the respective
realization conditions” [31, p. 5]. Put more perspicuously,
surefire dispositions follow the ‘realization principle’: if a trigger of
a disposition occurs, then does its realization [
31
]. Canonical
causation can be thus recast as the causal relation that is
grounded in some sure-fire disposition.
Not only canonical causation but also probabilistic causation
[
26
] is part and parcel of the biomedical practice, as is observed
by the ubiquity of probabilistic and statistical notions in the
medical domain: e.g., the probability of a person to contract a
disease within a given time frame. A straightforward
dispositional interpretation of probabilistic causation would
depend on probabilistic dispositions such as the disposition of a
coin to lands on heads when it is tossed up. Barton, Burgun and
Duvauferrier [
49
] elaborate upon probabilistic dispositions and
maintain that they fail to obey the realization principle because
the triggering process of a probabilistic disposition can occur
without its realization occurring.
All those findings show that probabilistic causation would be
explainable in terms of a direct extension of a dispositional
understanding of canonical causation based on probabilistic
dispositions. Consider the aforementioned paradigmatic example
of probabilistic causation: a flip of a coin caused the coin to land
on heads. In dispositional parlance, this was the case in virtue of
the fact that the process of the flip of the coin triggered the
probabilistic disposition of the coin to land on heads, which was
in turn realized in the process of the coin landing on heads. Note
that the probabilistic disposition of the coin to land on tails was
triggered at the same time, but it was not realized.
B)
Simultaneous Causation
As frequent in and as important for biomedicine as probabilistic
causation is simultaneous causation, where the cause occurs at
the same time as its effect [
50
]. Typical examples include a key
and a lock such that the former opens the latter. A certain lock
(say Lock1) becomes open only when it is unlocked by some key
(Key2) and vice versa. It may be tempting to speculate that the
Key2-pivoting-in-Lock1 process caused the
Lock1-opened-byKey2 process and vice versa. This would seem to go beyond
canonical causation because it fails to preserve the property of
asymmetricity. One may be also inclined to use the term
‘process’ with an emphasis on its dynamic connotation to
describe this apparently simultaneously causal scenario (see e.g.,
[
11,12
]).
We suggest that alleged simultaneous causation be explicable in
terms of the idea of reciprocal disposition [
51
], which can date
back to complementary dispositions [
43
] and the reciprocal
dependence among dispositions [
19,20
]. For instance, Lock1 has
the disposition d1 to be opened by Key2 and Key2 has the
disposition d2 to open this particular Lock1. Then d1 and d2 are
said to be reciprocal dispositions in that they can be triggered by
the same process universal and they can be realized in the same
process universal; and d1 bears the
has_reciprocal_disposition_of relation [
51
] towards d2. This
relation is irreflexive and symmetric. For instance, d1 (resp. d2)
is not reciprocal of itself and the has_reciprocal_disposition_of
relation holds between d2 and d1.
The notion of reciprocal disposition can be employed to capture
supposed simultaneous causation. In the case of lock-opening,
the Lock1-opened-by-Key2 process should be elucidated in such
a way that it refers to either the whole Key2-opening-Lock1
process or the Lock1-components-moving process. In virtue of
the causal efficacy of d1 and d2, the Key2-pivoting-in-Lock1
process indeed caused the Lock1-components-moving process
but not vice versa; and they are temporal parts of the same
Key2opening-Lock1 process. Generally speaking, simultaneous
causation is explicable in terms of the ‘causal dependence’ [19,
p. 5] between two reciprocal dispositions.
C)
Mental Causation
Mental and behavioral disorders constitute an acute problem for
the public health all over the world [
52
]; and biomedical
ontologies must take seriously mental functionings. An ontology
of mentality would facilitate an interdisciplinary research on
mental disease, thereby contributing to the improvement of
psychiatric diagnostics and treatment [
53
]. In particular, it would
help to fill a semantic gap between affective science and
psychiatry, which have been historically separate
notwithstanding their common goal to explore human mental
phenomena [
54
]. Several ontologies of mentality have been
built, including an OBO ontology the Mental Functioning
Ontology (MF) [
55
].
Mental causation [
56,57
] nonetheless remains elusive from an
ontological viewpoint despite its centrality to mentality. For
instance, it is fairly difficult to figure out what it is supposed to
mean to say: “I ate a cake because I was hungry.” To simplify
the matter, we assume the so-called belief-desire principle
according to which an agent’s action is basically a causal
consequence of her beliefs and desires. Inspired by Bratman
[
58
], the Belief-Desire-Intention (BDI) model of agency [
59
]
indeed recognizes the primacy of intentions (as well as beliefs
and desires) in practical reasoning and rational actions, and it is
widely used in applied ontology in virtue of its implementational
and logical benefits [
60,61
]. We omit to posit intention as a
fundamental mental entity, however, partly owing to
considerable controversy as to the (ir)reducibility of intention to
desire-belief pairs [
62,63
].
Seen dispositionally, mental causation would be most
straightforwardly explained in terms of dispositionality of both
belief and desire. As for belief, Barton, Duncan, Toyoshima and
Ethier [
64
] sketch out an ontology of belief in alignment with
MF and argue for two ontological meanings of the term ‘belief’:
a dispositional belief as a disposition that can be realized in an
occurrent belief and an occurrent belief as a MF-mental process
of taking something to be the case. A mental process is a bodily
process which brings into being, sustains or modifies a cognitive
representation or a behavior inducing state, where a behavior
inducing state is a bodily quality inhering in a mental
functioning related anatomical structure which leads to a
behavior of some specifi sort [
55
]. They say: “A dispositional
belief exists even when we are not actively thinking it, and when
we are actively thinking about a belief, we engage in an
occurrent belief process during which we take something to be
the case” [64, p. 4]. For the sake of our own terminology, we
will use the terms ‘belief disposition’ and ‘belief process’ to
refer to a dispositional belief and an occurrent belief,
respectively.
This Janus-faced view of belief can be extended to desire,
although its ontological nature has been little studied carefully in
applied ontology. Intuitively, desire is intimately linked with
motivation, which is in turn with behavioral dispositions. In
wanting to eat a cake, for instance, Mary is plausibly taken to be
disposed to act (e.g., to go towards a refrigerator) to satisfy her
hunger desire. As the most orthodox, motivational theory of
desire goes [
65
], for an agent to desire p is for the agent to be
disposed to take whichever action she believes will satisfy p (see
[
66
] for criticism). Based on this intuition and along Barton et
al.’s [
64
] line of argument, we can identify two ontological
interpretations of the term ‘desire’ and call them ‘desire
disposition’ and ‘desire process’. At first approximation, a desire
disposition is a disposition that can be realized in a desire
process; and a desire process is a mental process which modifies
a behavior inducing state so that the behavior inducing state will
lead to a behavior to satisfy the agent’s desire. Additionally, our
assumption of the belief-desire principle would imply that, to
realize itself, a desire disposition needs to be triggered (or
‘cotriggered’ if it is preferable) paradigmatically by multiple belief
processes.
Our dispositional picture of belief, desire, and action would be
still incomplete unless we incorporate into it an ontology of
plan(-making) because it is integral to practical rationality.
Following Barton et al.’s [
64
] recommendation, we extract from
an OBO ontology the Ontology for Biomedical Investigations
[
67
] the category of plan specification, which we here
reinterpret more broadly as a generically dependent continuant
which directs an agent (e.g., an experimenter) to perform a
certain action (e.g., an experiment). A generically dependent
continuant is a BFO-category: “A continuant that is dependent
on one or other independent continuants and can migrate from
one bearer to another through a process of copying. We can
think of generically dependent continuants as complex
continuant patterns either of the sort created by authors or
designers or (in the case of DNA sequences) brought into being
through the processes of evolution” [20, p. 179]. Examples
include the pdf file on Mary’s laptop and the pdf file that is a
copy thereof on John’s laptop. Characteristically, a generically
dependent continuants exists only if it is concretized in some
counterpart specifically dependent continuant. To take one
example, a paragraph of a novel in concretized in each pattern
(quality) of ink on the pages of the printed novel.
Mental causation is to be understood in terms of the application
of belief and desire dispositions (and their processes and related
plan specifications) to the schema of canonical causation. To
illustrate this point, imagine that a hungry girl Mary went
towards a refrigerator to eat a cake inside it. Our rough
dispositionally causal analysis of this scenario proceeds as
follows. First of all, Mary’s hungry physical process triggered
one (say dB) of Mary’s belief dispositions, which realized itself
in her belief process of taking to be the case that there is a cake
in the refrigerator. Next, this belief process triggered one (say
dD) of Mary’s desire dispositions, which was realized in her
desire process of wanting a cake. Most importantly, there is a
plan specification which is concretized in both dB and dD and
which induces Mary to approach the refrigerator to have a cake.
To satisfy her hunger, Mary might have taken another action
(e.g., to go to buy a cake) if she had believed otherwise: e.g.,
“There is no cake left in the refrigerator and a nearby cake shop
is still open.” It is interesting to remark that one may sometimes
use the term ‘intention’ to refer to a plan specification that is
concretized in belief and desire dispositions, and intention could
be therefore classified in our present framework as a subtype of
plan specification. Finally, Mary’s desire process under
discussion modified her behavior inducing state to make it lead
to her behavior of going towards the refrigerator to eat a cake,
hence the satisfaction of her hunger desire.
IV. Discussion
A)
Counterfactual
Let us consider briefly counterfactuals from the BFO viewpoint.
Counterfactuals are, broadly speaking, statements that represent
what did not happen or what is not the case: e.g., if a glass were
pressed with a certain force, the glass would break. It is typically
expressed by dint of a counterfactual conditional, i.e., a special
case of the subjunctive conditional, which uses what is known in
grammar as the ‘subjunctive mood’. A counterfactual
conditional has the form “If P had been the case, then Q would
have been the case” or “If P were the case, then Q would be the
case.” It is not hard to see the significance of counterfactuals for
biologists’ and medical professionals’ practice because scientific
procedures (e.g., experiments and clinical trials) generally take
on a crucial counterfactual aspect [
10
]. Clinicians would
ordinarily think, for instance: “If this patient had taken more
vitamin C, she would not have contracted scurvy.” Moreover, a
counterfactual analysis of dispositions is a well-known
traditional understanding of them and its core idea is that an
object is disposed to a realization r as a response to a stimulus s
iff the object would realize r if s were the case [
68
]. This view
has been subject to numerous critical appraisals, though [
69,70
].
We submit that, when grounded dispositionally, counterfactuals
are to be interpreted in terms of a generically dependent
continuant that is concretized in some disposition. For one thing,
counterfactuals are a kind of statements, which accord arguably
most suitably with the category of generically dependent
continuants (e.g., information). For another, the crux of the
dispositional grounding of counterfactuals is that counterfactual
conditionals are made true by dispositional ascriptions (but not
vice versa) [
31
] and this entailment would be ontologically well
explicable in terms of the concretization-in relation (which we
could reinterpret as one of the BFO family of grounding
relations) between a counterfactual (which we construe as a
generically dependent continuant) and some disposition.
B)
Explanation and Hypothesis
Ontology of natural necessity will exert a downstream effect on
epistemology of sciences, since scientific knowledge is about
scientific phenomena that are to be foundationally captured by
the notions of natural necessity. We will focus on scientific
explanation because it is of central importance to sciences
(including biomedicine), which consist in providing an adequate
explanation of the world. One paradigmatic example of
scientific explanation is that sky is blue because the molecules in
the atmosphere of the earth will scatter more blue light towards
the ground than other colors. In applied ontology, the notion of
explanation has been considered especially in connection with
observations [
71,72
], but it has been scarcely ever studied from
the perspective of natural necessity.
There are two important points which prior philosophical
models of scientific explanation have in common, although we
omit to delve into them owing to spatial limitations (see [
73,74
]
for a general survey). First, a strong link between scientific
explanation and causal explanation is generally agreed upon. It
is certainly debatable whether all scientific explanations involve
causal statements, but it can be safely said that scientific
explanation is paradigmatically underpinned by causal
explanation, as is implied by the ubiquity of the terms ‘causes’
and ‘because’ in the scientific literature. Second, and not
surprisingly, causation plays the most important role in
characterizing scientific explanation in the realm of natural
necessity. Each model is evaluated crucially with respect to its
ability to accommodate causal relevance. All this would
vindicate the extension of the grounding relations between the
ontological notions of natural necessity to the epistemic notion
of explanation: causation grounds causal explanation, which in
turn grounds scientific explanation, hence the causal grounding
of scientific explanation. For the sake of simplicity, we will
henceforth employ the term ‘explanation’ to refer to causal and
scientific explanation.
In our BFO framework, explanation is plausibly taken to a
subtype of Information Content Entity (ICE) from an OBO
ontology the Information Artifact Ontology [
75
]: a generically
dependent continuant of which that some material entity is a
bearer and that is about a ‘portion of reality’, which covers not
only BFO categories but also universals, relations, other ICEs
and configurations. An ICE is concretized by some information
quality entity as a subtype of quality. For instance, an ICE to the
effect that Donald Trump is the President of the U.S. is about the
configuration of Trump having the role of the U.S. President and
it is concretized in a string of alphabetical characters on this
paper. Given the dispositional grounding of causation,
explanation may well be seen as an ICE that is about some
disposition and that is generally recognized to be true in a given
scientific community. When biologists agree that an egg is
fertilized when it meets a sperm, for instance, an ICE to that
effect is about the disposition of the egg to welcome the sperm
(as well as the disposition of the sperm to enter the egg) and this
ICE is concretized in their mental qualities [
75
]. Along this line
of argument, we could ontologize (scientific) hypotheses,
although their accurate representation necessitates so careful
treatment within the BFO realist framework as to go beyond the
scope of our investigation (see e.g., [
76
]). Here we just say that a
hypothesis may be interpreted as an ICE that is about some
disposition and whose veracity has not yet been definitively
established within a given scientific group.
V.
Conclusion
In summary, we pondered over an ontology of causation in
relation with the biomedical domain, thereby arguing for the
dispositional grounding of causation and illustrating it with
canonical causation (and quasi-causation) in our terminology.
We then expanded our dispositional view of canonical causation
into other sorts of causation that are relevant to biomedical
practice: probabilistic causation, simultaneous causation, and
mental causation. We finally discussed a dispositional approach
to counterfactuals as well as scientifically important epistemic
notions such as explanations and hypotheses. Our findings can
be succinctly summarized in Table 1.
Future work includes the extension of our dispositional theory of
causation to type-level causation (i.e., the causal relation
between process universals) because ontological realism
compels ontologies to represent, inter alia, universals [
21
]. It
will serve as a useful starting point for our discussion to construe
type-level causation as a ‘causal law’ to be used for explaining
token-level causation [
9,11
]. Further investigation is also
warranted into our deeper dispositionally causal understanding
of other notions of natural necessity such as laws of nature and
epistemology of (biomedical) sciences, e.g., with respect to
observation and prediction.
Apply probabilistic dispositions to canonical
causation.
Do not take it at face value and focus on