Moral
particularism is a promising new approach which understands itself as a
subchapter of holism in the theory of reasons. So particularism may be extended
to other areas, such as metaphysics. One of the bases for this kind of move is
elaborated by particularism itself as resultance, a strategy for providing the
relevant basis that is opposed to various forms of generalism (the thin
property of goodness is constituted by several thick properties, such as being
good humoured, being pleasant; the property of this being a table is
constituted from properties of there being four legs, a plate, a certain
arrangement). It is claimed that resultance or emergence needs a background
structure in order to get off the ground.
Moral
particularism
Moral
particularism is an important new approach. The field of ethics and of moral
philosophy have undergone dramatic changes because of it. Ross and Moore get
widely discussed. There are books on particularism (Little-Hooker 2000) and by
particularists (Dancy 1993, 2000, forthcoming 2004).
What is moral particularism? It is a
view that there are no general moral principles guiding our actions, and that
moral action is based upon the insight into singular non-repeatable cases.
Moral decisions are lead by the normative authority of the particular, and not
by the normative authority of the general. Such a view is in disaccord with the
usual opinion according to which moral life and education has to be grounded
upon moral principles.
One main idea leading to
particularism is that there are several ways of organizing an area: that of the
exceptionless general rules, that of generalities which allow for exceptions
such as ceteris paribus clauses, and that of the singular cases whose list
cannot be subsumed under any general principles. Dancy’s moral particularism is
a byproduct of an endeavor starting in the area of causality. The main idea
there was that causality cannot be treated in atomistic manner coming under
jurisdiction of general exceptionless laws. The approach was eventually
elaborated for the area of morality.
An important ingredient of
particularism is that of the patterns that get involved into it. Generalists
think that the only metaphysically feasible and explanatory plausible items are
general patterns, which act as subsumptive devices.[1]
Particularist will point out the importance of singular patterns, i.e. of the
patterns characteristic for each single case, whose validity cannot be
generally extended. And, if it gets extended, then this is rather the matter of
a cosmic accident upon the canvass of the picture governed by the paint of the
singular.
Monism
and pluralism; generalism and particularism
In order to
see how it is possible to have a plausible view avoiding general patterns as decisive
items, it is instructive to have a look at generalist strategies of monism and
pluralism. Monism claims that just one general principle organizes the area of
morality, such as this is for example the principle of utility in the view of
the utilitarian. But besides to other flaws of such a position, it is not clear
how moral conflict would be possible under it: if there is just one principle
governing all of our moral decisions, why would one come into the situation of
indecision and of conflict?
Moral pluralism takes care of cases
of the now mentioned sort, affirming that there exists plurality of moral
principles and not just one. Sartre’s case of a young man is usually mentioned,
who has to decide between the duties of fighting Germans in the occupied
France, or of helping his old mother. There are two duties in conflict here,
one of the patriotic kind, and another involving respect for aging parents.
There is a room for a genuine conflict of duties. The conflict has to be
approached with insightful judgment encompassing the overall situation. Another
example figures yourself driving to an important meeting where you firmly
promised somebody to take part. While driving there, you encounter a person
involved into an accident, yourself being the only person around who is able to
provide the necessary help. You are thus torn between the duties of keeping
your promise, and of helping the person
in need. Again, you have to decide on the basis of your involved intuitions. If
you decide to help the person in need, you may feel regret, according to the
pluralist approach. This regret will be felt because one principle (that of
keeping the promises, in the discussed case) was not silenced, but was only
temporally put out of voice. There is plenty space for genuine regret in
particularism. So particularist will also have to account for regret felt in
the described case. But the regret will then not be related to the breaching of
general moral principles. Rather, it will be related to the impossibility to
take account of all the morally relevant properties and tendencies in a certain
particular case, and to such sources as that a certain action has failed to fit
into the track of one’s style of living. Particularist will thrive on
intuitions, but not on intuitions that would involve clash of principles.
Rather, there will be the intuition based on singular and unique patterns of
each involved case. Perhaps one can say that particularist will follow the
tendencies of the relevant and not of the principled regret.
Atomism
and holism
Atomism claims how in considering a
decision to undertake an action, such as a moral action, one has to take care
of a single feature, without considering its embedding into the context in
which it appears. It is then a natural suggestion that the support of such a
single feature will be measured by tractable and possibly exceptionless
procedures and by principles under which it falls.
Holism, on the other hand,
understands a feature as primarily coming embedded into a complex contextual
setting, in the framework of which only it is capable to exercise its
influence. The contribution of a feature to the overall situation will then not
possibly be measured by tractable means and exceptionless principles. Rather,
the insight into the overall situation will be of primary importance. Holism
offers a natural environment for particularist deliberation, whereas atomism
lies close to generalist techniques.
A wider
look at particularism: although particularism is formulated for the area of
morality, it may be extended to other areas, such as metaphysics. This
extension of particularism is a natural follow-up to the self-understanding of
particularism as a subchapter of holism in the theory of reasons
Particularism
was mainly formulated in the area of morality up to date. But the teaching of
primacy of normative authority for particular cases does not need to be
restricted to the area of morality. It is at least logically possible that
particularism may be elaborated for other areas such as metaphysics, say, or
for the area of epistemology.[2]
Of course the specificity of each of these areas should be then taken into
account. Dancy himself, as the main proponent of moral particularism, started
his deliberation in the area of causality, his first intuition being that
causality cannot be appropriately accounted for by the atomistic approach:
causes rather have to be portrayed in a holistic manner, without any general
exceptionless or even general principles with exceptions being involved into
specifying them. And accordingly, Dancy is sympathetic or at least he is not
contrary to the extension of particularist approach to other areas.
More generally, Dancy also
understands moral particularism as a subchapter of a wider approach of holism
in the theory of reasons. So reasons for action, and not just moral reasons,
are to be handled in a holistic manner. But this then invites the thought that
other areas may be approached in a holistic or in a particularist manner as
well.
Resultance:
a strategy for providing a relevant basis that is opposed to various forms of
generalism
One
important area in the particularist approach involves resultance, the feature
that accounts for the relation between moral properties and between the basis
upon which these moral properties get grounded.
But the idea of resultance is
rather wider and it extends way beyond the cases of morality. Where does the
property of this cliff being dangerous come from? It is grounded upon
properties of its being slippery and of its fragility. Where is the property of
this entity being a table actually grounded? It is based upon such properties
as there being four legs, a plate here, and a certain structural arrangement of
these. The property table results from or is constituted by these grounding
properties. It is not difficult to see that grounding properties may themselves
be further branched into sub-grounding properties, such as the property of the
plate resulting on properties of this material, of certain rigidity. The
property of this act being good results from its grounding properties of being
well-minded, of being humble, of happening or being acted at the appropriate
time and place.
The relation of resultance is
appropriate for a particularist in that it does not appeal to any general principles.
The grounding relation is not guided by any generalities that would contribute
to the result of goodness, say. Goodness is a thin property (thin, because it
does not have a lot of content characterizing it). A generalist will suppose
that thick properties (that of being well-minded, humble) always come with the
same polarity. This is countered by a particularist, who opposes the tendency
to see thick properties as instrumental only in respect to thin moral
properties. The particularist will try to relieve the stress from thin
properties, by emphasizing the shape brought in by the holistic environment of
intertwined thick properties and of the project of somebody’s own itinerary.
Resultance provides the possibility
of delivering a particularist account of grounding relation, which enables to
argue against various forms of generalism. Generalist approaches do not just
come in the shape of monism, but mostly in various forms of Rossian pluralism.
They all try to claim that a feature has the same contributory weight in all
the diversity of circumstances: if they have the positive weight in this case,
say, they will have to retain the same weight in all the subsequent cases.
Generalists will also tend to allow for ceteris paribus soft laws and thus for
exceptions, as long as the normative authority of the general will remain
unshattered.
Resultance provides one basis for
battling all these sorts of generalism. Resultance does not get the relevance
of features from the general rules being involved into these. Resultance
grounding somehow provides the relevance for free. Well, not entirely for free,
as we will try to indicate later on; there will be a non-classical structure
involved into the endeavor of its shape.
As far as the grounding relation is
concerned, resultance provides a basis that is quite different to that of
supervenience. Supervenience is the grounding relation account from the side of
the generalist strategy.[3]
Humean
argument that there can be no goodness besides to the thick good properties,
that there can be no mind besides to the thick mental properties.
Implausibility of this: compare: "there can be no table besides to the
thick properties of there being four legs, of there being a plate, of there
being a certain arrangement of these parts. You can perceive just these thick
poroperties, whereas table is nothing but an idea in your mind". This last
argument is clearly not plausible, and the reasoning may be extended back to
the implausibility of other Humean arguments. So, resultance or emergence is
possible (but as we will see emergence still needs a relevance encompassing
structure in order to be able to get off the ground).
There is a
consideration in the direction that relevance cannot even get off the ground.
This is a Humean consideration spreading skepticism over the very possibility
of properties resulting upon a diversified base. The strategy is skeptical as
to the transition from thick to the thin properties, and in this sense would
have to be close to some particularist efforts to relieve the stress off the
thin properties.[4] But again,
as already claimed, it is established first of all in order to argue against
the very possibility of resultant relations.
A Humean may say that there is no
property of goodness besides to the thick good properties of benevolece and of
helpfulness, in a certain case. Or that there is no property figuring the self
besides to the various thick mental properties. Wherever you will look in your
mind, you will only find some thoghts, or impressions, or again conscious
events, but no self. In the same sense you will find just thick moral
properties of benevolence and non-selfishness but no thin property of goodness
besides to these.
The arguing in this direction may be
easily shown as not appropriate on the basis of the following consideration.
There are these thick properties of there being four legs, of there being a
plate, of there being a certain arrangement. But there is no property of the
table resulting from these. The property of table would be just something that
we inappropriately construe in our minds. This reasoning seems utterly
implausible, as everyone easily sees that there is a table there. And so there
are good deeds and minds.
This establishes that resultance or
emergence is possible. I.e., emergent properties of goodness or of table may be
accounted for. But just how are they accounted for and just what brings their
relevance into the picture is another matter. It relates to the question of the
structure supporting the emergentist relevance.
Resultance
as following particularist unique patterns (beautiful patterns) as opposed to
the generalist strategy of supervenience. "Because" as an important
characterization of relevance. Did you help this person in this case because of
the general principle prescribing such a help to you? The answer rather seems
that you helped her because
of a rich variety of properties and characteristics involved into the
particular situation. This richness and dynamics of the forces involved into
the situation was the reason for you helping her, and not any general
principles.
Resultance
is distinguished against generalist strategies in that it does not claim how
the resultant property, in order to come about, would need any general
patterns. The resultance rather builds on unique singularly shaped patterns,
which may be called beautiful patterns in respect to the aesthetic pattern
structures that support the beauty of works of art. As opposed to this,
supervenience is the generalist strategy accounting for grounding relation,
which will be more closely dealt with in the following section.
One way to account for the
resultance is the characterization of “because” involved into it. Did[5]
you help this person now because of the general principle underlying
your activity? This may be what you claim. I.e., in providing the reason for
your action you will use the appeal to general principles: “I helped her
because one always needs to help people in need”. But you obviously do not help
all the people in need around the world during all the time. This would change
your life profoundly and it would not be feasible. You used the appeal to the
general principle because of the epistemic handiness it is dued with. In fact,
your reasons to help her in the situation that we are discussing were much more
diversified, depending both on the multiplicity of factors inolved and on your
ability of quickly discerning and deciding in which way it is proper to engage
in one’s actions. The richness and dynamics of the forces involved into the
situation was the real reason for your activity of helping her, and not any
general principles, to which you may appeal in an epistemic manner.
The
implausibility of the supervenient project: it is possible to construe a
generalist supervenient pattern just on the basis of implausible holistically
exact match of two situations. Objections to this strategy: The doubling of
twin situations is rather a cosmic accident (the utter improbability to have
two exactly similar St. Francis Twins). Supervenient relation tries to bring in
the guarantee for the general pattern to come through, without considering
anything else. But there is no relevance. Supervenient generalist strategy is
introduced by counterfactual lawlikeness.
Supervenient
grounding strategy opposes the resultance based or emergentist grounding
strategy. What does supervenience do? Certainly it is an account of grounding
relation. Supervenience will try to explain, as for that matter, how the
property of goodness will be grounded on a certain underlying basis.
The generality of supervenient
tactics may be seen because of its basic appeal to general principles or to
general laws, which are brought in by considering counterfactual situations.
Counterfactuals all by themselves support generalizations by securing law-like
connections. There are counterfactual lawful generalizations.
Take an example of supervenient
relation. There is St.Francis, and he is a good person. The property of
St.Francis being good supervenes upon his property of being humble, of his
helping the people at appropriate times and places. But this is not enough to
establish a supervenient relation. Supervenient relation needs a guarantee
of the generalist sort. This is why one introduces St.Francis twin, with
exaclty the same properties in its subvenient thick basis as the original
St.Francis. And here comes the main generalist induced claim concerning
counterfactuals and determining the supervenient property of goodness. If
St.Franics is good, and if there would be this St.Francis twin, wih the same
subvenient physical and thick moral properties as these that are found in the
original St.Francis, then St.Francis twin cannot fail to be good as well. This
is to say that the property of goodness supervenes on both occasions. But this
is then also determining the supervenient property of goodness itself, deriving
it from the generalist pattern according to which it is supposed to proceed.[6]
But consider now that all this
supervenient construction was actually made because of the wish to
deliver a generalist account of the situation. Just a moment’s thought will
reveal to you that the whole construction established in view of providing a guarantee
of the general was there just for this matter: to secure a generalist lawful
relation in order to support the supervenient property of goodness. Waiving
that, it must seem utterly implausible to have the situation of identical
St.Francis twins in the world. Could not already the original St.Francis be
considered as being good, without these generality lawful involving
considerations introducing supervenient properties? It is actually a
probability of accidence involving cosmic proportions to find two exactly
similar situations. But the adherent of supervenience has gone to big troubles
in order to secure such an unlikely case – obviously just in order to claim
that the property of goodness will have some authority only if it is
established upon some generalist pattern. But this presupposition (that there
are general patterns needed in order to act morally, say) is in no way to be
taken for granted. The particularist will claim just to the contrary that there
aren’t any generalities involved into a deed being good. It really depends on
this particular situation. Each feature may change its valence from one
situation to another one, from one context to another. There is no need for any
general pattern of this kind as the basis for having good acts, from the side
of particularist.
So this entire generalist
supervenient grounding strategy heads into a wrong, unrealistic and implausible
direction. Once one realizes this, the strategy proposed by resultance seems to
be a much more plausible alternative for an account of grounding relation.
There is no necessity for the appeal to the general rules. All that which is
needed is just the reliance of particularist unique and interestingly shaped
patterns involved into each particular context.
Resultance
is emergence: the basis thrives just upon the relevant properties, not upon all
of the properties as this is the case for supervenience. There is no relevance
in the supervenient basis.
Resultance
is actually emergence, as this was already hinted at. Dancy himslef called the
relation emergence, but he was persuaded to give up the offensive title,
offensive at least in the mind of his publisher, for resultance. Again, one
should consider that there is some history to the term of resultance. Ross
spoke about toti-resultant and parti-resultant properties, toti-resultance
dependning on the whole situation, and parti-resultance being grounded just in
a part of the relevant context.
Abolishing the recourse to any
general patterns enables one to establish the grounding resultance basis upon
the relevant properties exclusivelly, and not indiscriminately in its
dependency upon all the properties, as this was the case for supervenient
counterfactual relation.
There cannot be any relevance in the
subvenient basis, because all the physical and all the thick properties
are included into it. But relevance certainly needs some considerations of
salience, which are not provided by generalist supervenient strategies.
It is a great idea to have such a
relevant basis with resultance. But it will not come without some support of
the particular pattern involved into it, the pattern that is able to provide
relevance without any appeal to generality. Patterns involved into works of art
(pattern of this painting, pattern involved into the complex Don Juan opera)
are the cases in point. I think that Dancy should show the way towards such
patterns in order to be able to explain how relevance is capable coming off the
ground.
Dancy's
way of posing the dilemma: (1) Either there is generalism and atomism, general
principles and rule governed practices. (2) Or there is emergence, the relevant
properties. In this last case it is the matter of skills how to discriminate
properties. Comparison with the generalist semantics: there is no core meaning
of conjunction for the word "and"; the skill of discriminating the
semantic contribution of the word in varying situations is needed.
Dancy seems
to see the general dilemma involving particularism against the generalism
consideration just in the follwing way: either one goes with generalism or one
goes with particularism. But this is too short. There has to be a middle way of
a sort between these, and there needs to be some structure and some underlying
basis, a pattern that is not a generalist pattern. Dancy is too reluctant to
recognize such a structure, and so he fails to provide an appropriate account
of the particularist basis. Yet recently he has enabled a move in this
direction by opting for defaults and similar devices that should provide some
structure without that this would need to be an ultimately generalistically
based structure. One way to go here is obvious. There are default values of
cruel deeds say, which may turn out to be mostly wrong. But such a default may
be seen as an epistemic feature. So the default being just an epistemic
generality cannot really be a metaphysically grounded generality.[7]
A much more straightforward way is
to recognize the possibility of unique and nonrepeatable patterns, such as
various aesthetic patterns involved into works of art. This is the possibility
that does not seem to be recognizied to a sufficient extent by Dancy in
building a metaphysical underlying basis for his particularism.
Dancy seems rather to argue in the
following way:
(1)
Either
there is generalism and atomism, general principles and rule governed
practices.
(2)
Or,
there is particularism, emergence or resultance and relevant properties.
As he is
pointing to (2), Dancy depicts the matter of reaching the goodness of acts,
say, as being that of skills that are there in order to discriminate the
properties in a certain context.
But reference to skills somehow has
the undertone of there being no structure, as opposed to the generalist well
entrenched structure, proceeding over general patterns.
So, Dancy seems to miss the
possibility of there being a structure that is not a generalist structure.
He does argue in this direction to some extent in his seminal particularist
semantic sketch. The meaning of the world “and” does not have a core of
conjunction, according to this approach. There is no such core, rather the
matter of mastering the meaning of the word “and” depends on the skilfull
discriminating contribution of the word in the contexts in which it always
appears, such as in: “And what do you think you are doing?”
The opposition of full generalist
patterns supported rules and between the no rules but skills based practical
discriminatory ability is just too simplistic. There must be a more refined
possibility out there in the logical space: a relevance providing structure
without generality.
Helping
Dancy: providing him a way out about relevance. A structure is needed. What
kind of structure? We can go along with Dancy in acknowledging that it is wrong
to suppose just the existence of generalist repeatable patterns. There are also
the unique beautiful patterns. These are interesting patterns, they bear
relevance upon their sleeve: aesthetic patterns (the Gestalt pattern of Don
Juan opera, of Vltava symphony, of this picture as a work of art): they do not
repeat themselves, they are not included into any generalist pattern, but it is
because of this that they are the bearers of relevance. There is no relevance because of the generalist
patterns! And there is nothing more relevant than works of art. They give direction
to your life, if something does.
Let me come
again to the relevant patterns. These are patterns, obviously, not just
skilfull things without any real underlying structure. Skills should be
underpinned with some structure as well. Generative grammar for example is
understood as an underlying rich structure to linguistic skills. So in any way,
an account of the structure underlying skills is still needed.
A positive account of particularism,
not just arguing against various forms of genralism, is needed as well.
Phenomenology involved into action and into deliberating of action seems to offer
an obvious example. Delivering a viable example of non-generalist but relevance
providing structure is another one.
Aesthetic works of art seem to
provide the desired patterns. Interesting aesthetic patterns bear the relevance
upon their sleeve. Think on the unique and complex patterns underlying Don Juan
opera, or again on the patterns of Vltava symphony or again on those involved
into paintings in the gallery. All of these works of art are unique. They do
not repeat themselves. They are not included into any generalist pattern. If
somebody would write another Don Juan this would be just considered as a
plagiat or as some bloodless repetition of the primary powerful and not to be
repeated original.
On the other hand, aesthetic works
of art certainly seem to offer the prototype for bearers of relevance. There is
not any relevance in the works of art because of the generalist patterns that
may be involved into them. Even epistemic appeals to the generalist supported
relevance on their basis are more than dubious. But again, there does not seem
to be anything as relevant as the works of art. They may give a direction to
your life, they may help you in finding an unique path of your particular
journey.
The case
of relevant properties in resultance: thin property of goodness is constituted
by several thick properties, such as that of being good-humoured, being
pleasant. The property of this being a table is constituted by several thick
properties, such as there being four legs, a plate, there being a certain
arrangement.
Generalist
sees thick properties, such as these of being good humoured, being pleasant, as
always relevant and as relevant in the same manner at all occasions of its
appearance. Particularism, as already mentioned, will try to see thick properties
not as instrumental in obtaining the thin property, and as retaining their
valence through all occasions. It will rather see the thin property as not
being of any big importance. The reasons for action, say, will come from the
pattern of thick properties from such a particularist perspective, being guided
by the zero-level of itinerary, which may be described somehow as the way of
someone’s particular itinerary, the way someone actually leads one’s life.
There may be nothing dramatic in the way you live your life, but it will still
be a manner, a certain way in which you do live it. Some input of art will
certainly be able to provide a special quality to the way of your living your
life. The thin property is somehow not different from the thick properties that
constitute it; rather, it is identical to their arrangement.
Helping
Dancy: what is needed is something to support the emergentist structure of
resultance basis. What is thus needed here? A background structure!
The way of
helping Dancy is to supply something that will support the relevance. The
needed thing is the background structure. Emergence is not enough. Emergence
will not provide relevance by itself. A background structure will be able to do
this. One way to characterize this background structure is by invoking the
background, such as it was mentioned by Searle as he discussed intentionality.
Background is not the outright intentionality; it is the structure establishing
the preconditions of intentionality in an indirect way.
Example
of background underlying structure: morphological content. It is wrong to think
that there are just (1) general exceptionless rules, or (2) skillfully
accountable particular cases with no structure at all (compare Fodor vs.
Dreyfus). The logical space of possibilities needs to be extended. There is a
structure, which provides the relevance, the intractable background rich
structure, such as that of morphological content. It is wrong to think that
resultance/emergence does not need any structure. Example of background
structure for the case of morphological content: a rich multi-dimensional
landscape that operates without rules but that provides relevance with the
settling up of states into local minima. Emergence may not have and does need
an additional structure indeed, but this will not be structure based upon
general exceptionless rules.
Another way
to provide background structure is by invoking morphological content in the
area of cognition supporting the non-classical picture of models of mind.
Morphological content presents an example of beautiful pattern, i.e. of a
pattern that does not follow the requirements of generality in order to achieve
relevance. Classical models allow for three levels of description: a cognitive
system may be described at the levels of implementation, of algorithm and of
cognitive function. The description of implementation basically encompasses
physically realized hardware. Cognitive function centers at whatever is
outright displayed by the cognitive system, what may be interpreted as total
cognitive states such as beliefs or desires. Whereas the description at the
level of algorithm tries to capture the conditions for whatever is
displayed at the top level.
According to Marr’s classical
description, we find an algorithm at the cognitive system’s middle level. This
means that a kind of tractable procedure such as specified by some algorithm
assures the conditions for whatever appears at the top level of cognitive
system’s description. The conditions for this occurrent belief of mine to
appear are given by a tractable algorithm that may be specified, even if the
pattern it supports is a complex one, probabilistic or if it harbors some other
kind of diversity. The idea behind the algorithmic approach is that even for
such complex cases a tractable algorithm may be found that specifies their
general repeatable patterns as the conditions for what is displayed by the
cognitive system at its top level.
Horgan and Tienson (1996) propose a
substantial qualitative extension of levels involved into the description of
cognitive systems. The basic move happens at the middle level of cognitive
system’s description[8],
where “mathematical-state transitions”[9]
substitute the classical algorithm. What is going on here? The extended generic
model is inspired by connectionist models of mind, as opposed to the
classical language of thought based models of mind. Connectionist systems
proposed an architecture that is not based upon classical algorithms, but
rather upon dynamical mathematics inspired procedures. A rich multi-dimensional
landscape that operates without rules but that provides relevance with the
settling up of the system’s states into local minima underlies the endeavor.
By this move, connectionism certainly has shown the way out of the classical
architecture. But just adopting connectionist mathematical-state transitions
would be too short if the aim is providing a description adequate to the
performance of the actual cognitive systems. It would be too short if
connectionist inspired description would still retain as its departure some
tractable procedures, of a probabilistic or some similar kind. And it would be
too short if any claim in favor of the structure would be abandoned and if one
would give free course to a kind of computational anarchy. Connectionism is
just inspiration and entrance into deciphering of a non-classical relevance
providing structure, into a non-classical language of thought. The support for
such a structure comes at the middle level of the non-classical or generic description
as the structure of morphological content that supports
mathematical-state transitions. Without morphological content, these
transitions just are not relevant. In connectionist models of mind, as opposed
to the classical models of mind, the outcome gets determined by its’
positioning in the multi-dimensional mathematical space and not by any
algorithm.
What is morphological content
thus? It is the background content or background structure at the middle level
of non-classical generic cognitive system’s description, a structure that may
not be described by any tractable means, but that provides relevance or in
other words condition for whatever appears at the top level of cognitive
system’s description. Morphological systems may be described as an intractable
background landscape that provides the points for positioning of top total
cognitive states and that determines cognitive-state transitions. The structure
of the cognitive system at this middle level of description is intractable but
relevant for determining whatever appears at the higher computational level.
The relevance has to be obtained by something. Morphological content
does not provide relevance by repeatable or tractable general patterns but by
its holistic intractable unique structure. Notice, by the way, that this unique
particular structure is not static, but that it traces the relevant
points of positioning through time. This means that the relevance tracking
structure itself is substantially dynamic. Morphological content as the
background structure gets displayed at such occasions, as is the one where my
specific intonation accompanies my pronouncing of the English text, myself
being brought up as a speaker of Slovene language. More simply and obviously,
morphological content is active in positioning of anything that gets
occurrently displayed by my cognitive system: my thoughts, the things I say,
the acts I perform. The background structure determines all of these and most
of them are relevant for situations and activities I get involved in. For each
singular case, there is this dynamical relevant structure. Morphological
structure provides singular beautiful but dynamically changing and adjusting
pattern.
While discussing situation in the
area of models of mind, Horgan and Tienson (1996) assert the need for a broader
picture out there as just the one described either by exceptionless general
rules based models of mind on the one hand, such as Fodors’s classical language
of thought model, and between the connectionist models based conceptions of
mind, such as the ones promoted by skills (Dreyfus) and with absence of general
structure (Churchlands). There is a range of possibilites, which may be
inspired by connectionism, but which do not reject all forms of structure along
with the rejection of the generalist based structure. Such a possibility
involves what Horgan and Tienson call the model of dynamical cognition.
Dynamical cognition leans heavily on the idea of rich background virtual space
that shapes in many directions the way our cognition goes. This virtual
multi-dimensional space includes morphological content, which is not occurrent
content, but rather whatever is involved into the weights, if one may use
connectionism inspired terminology. Morphological content comes from the underlying
structure that shapes the direction of the forms of cognition to occurr.
In classical computational models,
whatever appears at the occurrent level of the system gets determined by the
underlying exceptionless structure or algorithm, by a language of thought.
Dynamical cognition, to the contrary, builds upon unique singular beautiful
patterns. These patterns still provide
a language of thought, but a non-classical one. This is a dynamics based
language of thought without algorithms.
Resultance or emergence needs an
underlying structure in order to have relevance embedded into it. If such a
structure as is the one brought by the morphological content is provided, then
we can escape the simplistic opposition between just (1) general exceptionless rules
on the one hand and between the (2) skillfully accountable particular cases
with no structure at all on the other hand. This would then be similar as the
opposition between Fodor and Dreyus in the area of models of mind (general
exceptionless atomistic rules as against skills that are nor based on any
structure, the no-structure here being understood as skills). The logical space
of possibilities has to be extended in respect to these. There is a structure
that provides the relevance, the intractable background rich structure, such as
the structure proper to the morphological content. Resultance or emergence does
need some structure indeed in order to be able to produce relevance. The
background structure of the morphological content may provide a guidance here.
Morphological content involves a rich
multi-dimensional landscape that operates without rules and provides relevance
with the settling up of virtual states into local minima. Emergence does not
have and it needs a structure indeed, but this will not be a structure that
involves general exceptionless rules.
Relevance
and the frame problem (a robot will not be able to act because of the
impossibility to get around on the basis of tractable rules). The action iself
is possible upon the rich and dynamical background of contributing and enabling
conditions as reasons for action, based on facts and context and not on beliefs
and desires (The reason for me to having helped her is the fact that she needs
help, not my belief in this direction; ultimately the fact as a reason may have
some weight in the context only).
One example
illustrating relevance involves the frame problem. Dennett describes a robot
with a bomb attached to it that will explode in a due time. Robot is also
equipped with a powerful classical computer based on rule governed symbol
crunching and with input devices of visual and other appropriate sorts. The
task of the robot is evidently that of recognizing that there is a bomb
attached to it, that it is in danger and that it has to act accordingly so that
it can escape the dangerous fate. The lesson of the frame problem is that there
are no chances for the robot to be successful if it follows the elaboration of
the input information according to the classical rule governed symbol-crunching
computer. The robot will look at you, it will elaborate the information related
to your eyes, to the color of your shirt. Well before it will be able to
realize that it might be in danger it will be too latte for it to undertake any
action for saving itself. The reason is that tractable rules will not by
themselves assure any relevance.
The action does not follow tractable
computing crunching, but it is rather possible upon the background of rich and
dynamical contributing and enabling conditions that figure as reasons for
action.
Dancy has criticized the view that
mental states, such as desires or beliefs, may figure as the relevant reasons
for action. If some of these would be able to figure as reasons, then this
would be beliefs because of their bigger amount of objectivity as compared to
the desires preferred by Humeans. But facts are even more objective. I helped
her not because I believed that I need to help her; the real good reason for my
action is the fact that she was in need, not my belief to this effect. But
there is something that is even more objective than the fact: it is the rich
context that provides an appropriate reason for my action of helping her.
The dilemma
for the particularist profiled itself in the following terms: either on the one
hand there are generalities that are responsible for moral or for other higher
order properties, or on the other hand there is just skilful underpinning of
particular cases. What are the skills we are talking about? One area for skills
is our mastering of language. What conditions are needed in order for someone
to understand a certain word? The meaning of a word is not mastered by
elaborating upon some kind of its core meaning, but by appropriately handling
the usage of the word in various contextual settings. The meaning of the word
“and” does not equal some general semantic core meaning such as conjunction.
For there are other forms of the usage of “and”, such as it appears in the
sentence “And what do you think you are doing?” If one agrees not to buy
generalities and if one embraces skills, there is still the question about the
structure that conveys relevance. Up till now we have not presupposed any
background structure to be involved into the skillful handling of cases. But
such a structure may perhaps be found if we look at the ways in which
resultance itself is shaped.
The following schema may be
construed along the lines of Dancy’s (1993) presentation of resultance:
0 – 0 -0
0 I 0
o -constituency→ thin moral property
thick moral
properties
Ý resultance
natural
properties
A structure
providing relevance may be found at two stages of the schema. These stages have
different tasks to fulfill. But taken together they also constitute an account
of a unique phenomenon. Schema may be illustrated by substituting “good” for
the thin moral property, “benevolence”, “friendliness”, “helpfulness” for thick
moral properties and whatever physical material and other properties underlie
these for natural properties. The goodness of my act is constituted by its
benevolence, friendliness and helpfulness, as well as by the ways in which
these are distributed in the context of moral situation. These thick properties
and their arrangement though result from the involved natural properties.
There are two relations involved
into the phenomenon in its entirety. The first is the relation of resultance.
The second relation of constituency does not proceed between two levels; it
rather succeeds at one single level. There are no generalities involved into
these relations at any of the involved stages.
The relation of resultance is that
between natural properties and between thick properties. The additional
relation of constituency succeeds between thick properties and between the thin
property. If there is thin property of being good, it is constituted by thick
properties of benevolence, of temperance, of efficacy, of friendliness and by
the shape of a certain situation in which these thick properties appear.
Constituency is not a mediating kind of relation. The thick properties
involved, such as friendliness and efficacy, together with their overall shape
in a certain case do not really appear at another level, as is the level of the
thin property that they constitute. Thick properties and their arrangement are
in fact the resulting thin relation. There is nothing else as these thin
properties and their arrangement or their form that would constitute the thin
property of goodness – in this particular case.
But thick properties, such as
friendliness and benevolence, result from other properties. So there is a more
indirect relation in the case of resultance as there is in the case of
constituency. Thick properties are grounded in and they result from natural
properties. But there are several layers of resultant properties involved into
the overall relation of resultance. The thick property of efficacy, for
example, results upon several properties such as keeping promises, doing
appropriate things, which again result from other properties underlying them.
Finally, at some stage, the resultant tree of such properties reaches the
natural basis. While there is such a resultant tree for the relation of
resultance, there is nothing comparable for constitution: thick properties
simply constitute the thin property. Thick properties and their overall
arrangement in a certain situation together just are the thin property.
In the area of artifacts, the table as the thin property gets constituted from
thick properties of there being four legs, of there being a plate and certain
arrangement in which these fit together. Table is nothing else as the
appropriate arrangement of these thick properties. But thin properties result
from a tree arrangement of their underlying thick properties (the table plate
results from this material of a certain density) and ultimately from their
underlying natural properties. The table is constituted by all of these.
Considering these two stages of the
phenomenon in its entirety, we may see that there is a structure providing
relevance involved into both of them. There is the structure of the resultance
tree and there is the structure of constitution. So we are not just talking
about the absence of structure, meaning thereby the absence of a general
structure in the case of resultance. To the contrary, resultance introduces a
rich and powerful underlying structure conveying relevance to whatever emerges
as its product.
Perhaps the phenomenon of bringing
resultance and constituency together repeats itself at several intermediate
stages of the resultance tree. Resultance tree is where thick properties result
from several underlying properties. The beginning of the branching is at the
level of the natural properties. But resultance tree usually contains a lot of
properties that are of higher level, as are the natural physical properties. So
it is possible to consider relation of constituency appearing already at the
level where the resultance operates.
We now wish to take a slightly closer
look at the relation of constituency. The question is about the kind of basis
upon which the thin property gets constituted. There are two candidate
possibilities:
(a)
thick
properties –constitution→ thin moral property
(b)
thick
properties + their form –constitution→ thin moral property
A
particularist will go for (b), which corresponds to the above schema figuring
resultance. If he would go for (a), he would finish with kind of type-type
constitution theory. This would claim that a certain type of thick properties
is bound to constitute a certain type of thin property. The adherence to the
type would bring generalities into the picture. But because of the variable and
intractable richness of the thick properties available to support the thin
property of goodness, say, it is questionable whether any such general type
would be forthcoming. It is also questionable to talk about the thin moral
property as about a type, because there is just one such property available,
goodness in our case. And because there will be no inherent form of the
constitution involved in (a), this proposal will be atomistic and thus it will
lack holism as the distinguishing mark of the particularist approach. This
would also not be compatible with the main claims of particularistic holism,
because types involve generalities: a certain kind of type has general and
lawfully predictable consequences for several cases. We then finish in an
atomistic kind of constitution relation – according to atomism a feature or a
certain set of features has the same consequences without respect to the
situation involved – which is incompatible with particularist holism.[10]
If we embrace the possibility (b)
though, the relation of constitution will be that of token-token identity. This
is a more natural move for a particularist. The form or shape of thick
properties coming together that is to be found in this case makes for all the
difference. We are talking about a holistic form, in the sense in which
a particularist understands holism. This means that there are these thick
properties. But as there is also the landscape together with its shape in which
these properties appear, their valence may change along with the altering of
landscape’s form. One such example would figure Dancy’s story about his young
daughter stepping on a sea urchin. The extraction of the urchin’s needle caused
pain, and causing pain is not a good making property. But because of the
overall situation in which this feature appeared, i.e. because of the shape
into which the thick properties came in this case, the contribution and the
valence of the feature changed. The addition of the form to the assemblage of
thick properties in (b) thus exercises holistic effect – the claim that the
valence of a feature may change in respect to the form of the involved context.
This makes the approach in (b) compatible with token-token identity theory.
Token-token identity namely does not commit itself to the generality of types,
it just claims in favor of a support that may be different for each particular
case involved. This also accords with the observation that there is an immense
variability and richness of thick properties underlying a certain thin
property. There is a practically infinite number of thick property bases for a
certain thin property such as goodness, with each of the thick characteristics
involved being capable of changing its valence in respect to the holistic form
of context in which it appears as the support for a given thin property. We
cannot construe any generalities in respect to how these relations get linked
between themselves. Now it is obvious why the form or the shape in (b) is
important in order that the approach can have a particularist touch to it.
A distinguishing characteristic of
the model according to (b) is that the form or the overall shape pertaining to
a certain case gets involved into it. Because of the adherence of the form to
the particular case no general predictions about how one may compute by
tractable means the result of several thick properties coming together at a
certain landscape may be made. We have to do with particular and not with
general patterns.[11]
This is also why there are no definitions really available about how to realize
necessary and sufficient conditions for good properties. The resultant property
just emerges upon a rich background that cannot be mastered by tractable means.
Such a resultant property[12]
brings the emergence based upon the particular pattern along with it.
There will be relevance achieved upon the form of such intractable
basis. And it will not be generalistically justifiable relevance but a particularist
relevance, i.e. the relevance achieved on the basis of particularly shaped
patterns. It now becomes doubtful whether relevance may be achieved upon an
underlying set of properties, without inherently coming in the shape of
particular unique patterns or forms. In one way, one may talk about the
property of goodness as resulting from a natural basis. But it also gets
constituted from the thick properties.
In fact, breaking down of the
relation of resultance and emergence has shown us a way out of the too
simplistic dilemma concerning a choice between general principles and between
the skills that do not involve any structure. The form of thick properties or
their overall shape from which a thin property is constituted provides the
needed structure. Just that this structure cannot be generalized; it is
a particular structure, emerging from each case. This is a good point
for one to stress that resultance should not be confused for supervenience.
Whereas supervenience builds upon general patterns or general forms,
presupposing that only generalities can provide a relevant structure,
resultance thrives upon singular patterns. Whereas relevance is inherent to
each singular pattern (consider work of art as an example, or your singular way
of living your life), the relevance involved into supervenience comes from
repetition of a certain general pattern. But it is hard to see how just a
repetition of a certain situation would bring relevance with it, except in
fallacious ways.
Generalist will try to argue that
relevance is to be achieved just on the basis of the general. There is this
thin moral property, such as good. But the thin property of goodness has many
instances, for there are many good acts around. But if there are many good
properties, they must have something in common – being good as for that matter.
But then, there obviously exists a generality or a general mechanism that binds
all of these instances. What can a particularist say at this point? He may
shift attention to the fact that a powerful illusion is exercised here,
a drive to recognize general patterns behind various instances. This drive is
fuelled by confusion between the epistemic and between the metaphysical.
Perhaps similarity and the general are features of the epistemic
assessment of the world. The epistemic assessment, such as conceptual
assessment, is driven by reduction of the richness and by recognition of
several common traits to a range of cases. But it is not the case that these
generalities would have a metaphysical underlying basis. Generalities,
from this point of view, are just something that gets projected upon the world,
with the means proper to language and thought, but without any real
metaphysical support.
Supervenience as a generalist strategy takes all
natural properties into account, and it does not mention any structure, such as
resultance tree. Once all properties are taken into consideration, it only
seems that one can achieve relevance on the basis of the repetition of complete
situations. It seems that by this means one may get to the general pattern. But
it is questionable that general pattern would lead to relevance.[13]
Because of this lack of the structure supporting it in the background,
supervenience does not thrive on relevance. For the case of resultance however,
the particular shape and the lack of definition leads to non-tractability and
to emergence, and therewith to relevance. Resultance recognizes that relevance
does not come from a general repeatable structure but from an underling
particular structure. [14]
Churchland,
Paul (1988). Matter and Consciousness. MIT Press.
Dancy,
Jonathan (2000). Practical Reality. Oxford: Clarendon Press.
Dancy,
Jonathan (1993). Moral Reasons. Oxford: Blackwell.
Dancy,
Jonathan (Forthcoming). Ethics
Without Principles.
Dreyfus,
Hubert (1979). What Computers Can’t Do. Harper and Row.
Fodor,
Jerry (1981). RePresentations: Philosophical Essays on the Foundations of
Cognitive Science. MIT Press.
Horgan,
Terence and Tienson, John (1996). Connectionism and the Philosophy of
Psychology. MA: MIT
Press.
Little,
Margaret and Hooker, Brad (2001). Moral Particularism. Oxford: OUP.
Potrč,
Matjaž (1999) “Morphological Content”, in Connectionism and the Philosophy of Psychology, Acta Analytica 22: 133-149.
Searle,
John (1983). Intentionality: An Essay in the Philosophy of Mind.
Cambridge: Cambridge University Press.
Verdiglione,
Armando (2003). Artisti. Milano: Spirali.
[1] In opposition to generalists, particularists think that general patterns (a) proceed without the real need of any ultimate material hookup (supervenience as a generalist strategy involves counterfactual comparing of several situations, without that any of these would have the natural or physical instantiation secured – physical realization is just an additional and even not necessary assumption in the strategy pertaining to the generalist counterfactual patterns); (b) are not explanatory for the areas in which they are often used, such as the area of morality: morality does not require explanation that is shaped according to the requirements of natural sciences including general laws; in the area of morality a suitable explanation would be that of narrativity, which is more appropriate for handling of particular cases, and gets rooted in non-repeatable but relevant patterns.
If particularism's metaphysical approach uses resultance or emergence, then this involves natural or physical basis as a more binding presupposition. And then narration seems to be an appropriate explanatory strategy.
[2] Lately, Dancy elaborated a basis for a particularist conception of meaning. His general idea was, as he now remarks, »that there should be an analogue to particularism wherever there is rationality to be found«.
[3] Dancy says in his commentary: »I wonder, at the end, whether it is true that supervenience is the generalist's form of the grounding relation. I know that many people think of supervenience as a relation of 'fixing': the subvenient base, in all its enormity, fixes the supervenient property. I myself don't think that this is the right way to think about supervenience at all, however. I would have thought that the generalist analogue for the grounding relation is more likely to be subsumption.«
[4] There is confusion in the Humean approach between the level of the property of there being a table and between the level of properties such as there being four legs, a plate, their arrangement, that constitute the resulting property of there being a table. The property table does not have the same function as the property of there being four legs. The property of there being four legs contributes to the constitution of the property of there being a table. But all the mentioned thick properties (there being four legs, a plate, their arrangement) also are the property of there being a table, in the sense that they constitute this property.
[5] Dancy's commentary: »The question should not be whether you did help the person because of a principle or because of the properties of the case, but whether the reason why you should have helped her was the principle or the properties. That is to say, I think it unwise to go off into the question what actually motivates people, and better to stick to the explicitly normative question about what grounds the relevant ought.«
[6] It is questionable whether any relevance may be derived upon the mere establishing of generalist patterns. Such a presupposition is widely accepted and nevertheless unfounded. Slightly more about this in what follows. Yet the matter would still merit a separate treatment.
[7] As Dancy rightly remarks in his commentary, »To think of something as a default is to think more generally about the way in which it functions in a variety of cases, rather than directly about how it is functioning in some particular case before one.« This is clearly a remark about the constitution. But if deafults are taken epistemically as I propose, they would at least be compatible with particular patterns. These would include relevance in several of its forms of salience and shape, of the contributory, the enabling and the intensifying conditions.
[8] The bottom level of cognitive system's description may stay the same for both classical and extended model. The top level of cognitive system's description introduces cognitive-state transitions instead of more static classical cognitive function.
[9] Horgan, T. and Tienson, J. (1996), p. 45.
[10] Type-type identity is not an unusual proposal in the theory of mind. If it just means something such as that
a type of mental property results from the natural, i.e. physical properties,
this is a comparatively harmless claim. Still, because of adherence of the type
identity theory to generalities it will be difficult to incorporate relevance
into it.
[11] It is essential to grasp that particular patterns are still patterns. I.e. they have relevance embedded into themselves, they have a structure, but they certainly are not any general patterns. So particular patterns prove that at least, not all patterns are general. Besides to this, particular patterns possess relevance, whereas this is not so clear for general patterns. I am enjoying this performance of this symphony now, not some general kind of performance. Works of art are relevant in their singularity. Giving examples helps the teaching, whereas appealing to generalities is much less certain outcome.
[12] One would perhaps be better off if talking about the constituted property. But as the overall phenomenon of resultance incorporates both resultant and constituing stages (these may be separated by analysis only, and not in reality), we may also talk about resultant properties.
[13] Compare general and practical syllogisms. »All people are mortal. Socrates is a man. So Socrates is mortal.« does not automatically have relevance for any specific instance. Whereas »If somebody is hungry, they will eat a cake. Here is a cake. I am hungry. So I eat this cake.« applies to particular cases and brings the relevance along with it. Notice that indexicals appear in this later case.
[14] My thanks in preparing of this
paper go to Jonathan Dancy for his discussion at the occasion of his Valencia
conference in March 2003, and for the written commentary concerning the present
paper.