An improved procedure for 2-amino-5-nitro-4,6-diarylcyclohex-1-ene-1,3,3- tricar bonitriles; carbonate on polymer support as mild and reusable catalyst

Article abstract of DOI:10.5012/jkcs.2011.55.5.808

A new catalytic system has been developed in the synthesis of 2-amino-5-nitro-4,6-diarylcyclohex-1-ene-1,3,3-tricarbonitriles using carbonate on polymer support (Amberlyst A-26 NaCO3 -). Short reaction time, simplicity of isolation, safe catalyst and high

Full text of DOI:10.5012/jkcs.2011.55.5.808

Mechanisms and Resources∗
Colin Klein
Macquarie University
cvklein@gmail.com
1 Introduction
Mechanistic accounts of explanation are popular (Machamer et al., 2000;
Craver, 2007; Bechtel, 2008). They are also controversial. Most of the
debate has been over whether mechanistic accounts are sufficiently compre-
hensive — that is, whether they cover all the explanations given in a partic-
ular domain. Everyone agrees that some explanations (especially in neuro-
science and molecular biology) are mechanistic. Explanations in chemistry,
or physics, or evolutionary biology are less obviously mechanistic, however.
Even within neuroscience there are explanations that don’t look terribly
mechanistic: those which appeal to general network properties, or to dy-
namic laws, or to computational processes, all present special challenges
(Stepp et al., 2011; Piccinini and Craver, 2011; Kaplan and Craver, 2011;
Levy and Bechtel, 2013). In the background, there is also a general worry
about whether mechanists properly distinguish models of mechanisms from
descriptions of those models (Klein, 2012; Craver and Darden, 2013).
All of these are interesting, big-picture worries. I want to put them to one
side and focus on a second kind of question. That’s the issue of whether ex-
isting mechanistic accounts are complete – that is, whether they say all there
is to say about explanations in the obviously mechanistic realms. Take some
paradigmatic mechanistic explanations: explaining the action potential, or
how an engine works, or how the kidneys do what they do. You might ask:
do mechanistic accounts of explanation capture everything going on here?
Do these explanations constitutively involve what the mechanists say they
do? Or is there something more going on?
∗Draft 1De1. Compiled March 2, 2015
1

Completeness hasn’t received nearly as much attention comprehensiveness.
Professionally speaking, it doesn’t seem like a sexy question. Charitable
readers recognize that all accounts are incomplete. Filling in the details is
the sort of cleanup job usually left for seminar papers, or done en passant
on the way to some more interesting conclusion.
That’s a shame, though. For I think that standard mechanistic accounts are
incomplete, and importantly so. I will argue that not all mechanistic expla-
nation involves citing components of mechanisms in the explanans. Some
mechanistic explanations involve citing resources that are used by mecha-
nisms. Resources can, and should, be distinguished from the mechanistic
components that use them. That’s not a mere tweak to the story.
I say more. One of the more persistent worries about mechanisms—present
in many of the more daring critiques above, and a source of grumbling in
conference hallways—is why we ought to be mechanists about explanation,
as opposed to adopting a more general causal or interventionist story. Even
if neuroscientists do give lots of mechanistic explanations, it might be that
this is just an especially compact way of giving causal information.
Resource explanations shed light on this question as well. They’ll do so by a
rather indirect route: having made the distinction, I’ll then argue that that
the mechanism-resource distinction is itself not exhaustive. Instead, they
are the most obvious results of an explanatory strategy that depends on the
interests of particular disciplines. The interests upon which that categoriza-
tion depends, though, are pretty important ones. Arguably, they’re the sort
that are constitutive of a scientific discipline. Because of that, the choice
of mechanistic explanations is not mere convenience, but constitutes a vital
explanatory strategy in the sciences that employ it.
2 Mechanistic Explanation
2.1 Parts and Wholes
I will focus on the account of mechanistic explanation given by Carl Craver
in his (2007). A mechanism, in Craver’s formulation, is “a set of entities and
activities organized such that they exhibit the phenomenon to be explained.”
(p5). Mechanistic explanations are fundamentally multi-level. They work
by demonstrating that the behavior of a whole can be captured in terms of
the organized activities of its parts (p139).
2

Suppose I want to know how a car engine works—how it takes in fuel and air
and uses them to produce radial motion in the crankshaft. A mechanistic
explanation will break the engine down into components, explain what each
component does, and show how it is related to the others. Having done
so, I can show how the organized activities of the components produce the
activity of the whole. I can show how the fuel rail is connected to the
injectors, and delivers fuel to them, and then the injectors spray the fuel
into the combustion chamber, and on and on. Once I have exhibited the
activities of all of the components, one can see how the engine as a whole
works.
Craver shows that this style of explanation is also found in many sciences.
To explain the action potential, we show the channels embedded in the cell
membrane, we explain the conditions under which those channels open, and
so on.
A crucial aspect of mechanistic explanation is breaking whole objects down
into their mechanistic components. A good account of mechanism ought to
give a criterion that distinguishes true mechanistic components and their
activities from mere spatiotemporal parts. If a wombat crawls into my
engine compartment and goes to sleep, that doesn’t make it a mechanistic
component of my engine—even if, note, it might have a variety of causal
interactions with the engine.
Craver appeals to the relationship of constitutive relevance in order to dis-
tinguish the true mechanistic components. Constitutive relevant is in turn
cashed out in terms of mutual manipulability. Some proper part x of S is
constitutively relevant to Ss ψ-ing if there is an intervention on xs φ-ing
that would change Ss ψ-ing, and if there is an intervention on Ss ψ-ing that
would change xs φ-ing (p. 153). In other words, it must be that one is able
to reliably manipulate the behaviour of a constitutively relevant component
to manipulate the behavior of the whole, and vice-versa. The behavior of
the fuel injectors can be manipulated by speeding up the engine, and the
engine can be manipulated by changing the speed of the injectors; that is
why the injectors count as constitutively relevant. By contrast, there’s no
(reliable) change I can make in the interloping wombat by manipulating the
engine, and poking the wombat has an inconsistent effect on the engine.
2.2 Abstracting Away
Mechanistic components are a diverse lot. Fuel injectors, hearts, sodium
channels, and syntactic modules appear to have, on the face of it, little in
3

common—other than that they are all cited in mechanistic explanations.
Abstracting away a bit, though, we can see that mechanistic components
share several important features.
First, mechanistic components persist over relatively long timescales — for
now, let’s say over the timescale of the behavior of the whole that we wish
to explain. The different engine components are present the entire time
the engine is running. That is an important feature of the mechanistic
explanation: we implicitly assume that when the fuel rail sends gas to the
injectors, the injectors are always there and ready to receive it.
Conversely, the failure of a mechanistic component is often the explanation
for the failure of the whole mechanism to perform. The Challenger exploded
rather than went into space because an O-ring failed. Tetrodotoxin kills
by blocking the activity of sodium channels. Mechanistic components must
thus be available and functional throughout the timescale of the explanation.
Failure to do so explains deviations from the norm.
Second, mechanistic components are individually important. In addition to
the availability of component-types, the identity of a component over time is
often crucial to the explanation. My car starts hard in the morning. Why?
It is too old to take ethanol-augmented gasoline. The ethanol destroyed the
check valve in the fuel rail. As it sits overnight it loses pressure and makes
for a hard start. Note that for this explanation to work, it is important that
the same entity was affected by the ethanol, leaks overnight, and therefore
explains what happens the next day.1
Thus mechanistic components don’t just persist, but can persist as the very
same entity over a variety of activities. This is obscured when we give
mechanistic explanations because we often care about explaining types of
mechanisms rather than individual tokens. But for any individual token
mechanism, it is usually assumed that the same components persist, and
persist as the same components, over time.
Third, mechanistic components are typically casually conservative. That is,
each component interacts with only a limited subset of the other compo-
nents in a small number of ways. The fuel rail does one thing, the intake
another, the injectors a third; further, the fuel rail interacts only with the
fuel line and the injectors, the injectors with the rail and the chamber, and
so on. Casual conservativeness allows for modularity, which is a feature of
complex mechanisms (Simon, 1996; Bechtel and Richardson, 2010). Further,
mechanistic decomposition is made possible by the causal conservativeness
1Historical footnote: it was actually the starter.
4

of the individual components. Most of the processes we care about are com-
plicated. We explain a complicated process by looking at the components
involved and the limited sets of relationships between them. Note that the
components need not be simpler than the whole. Rube Goldberg machines
show that it’s possible to build up a simple process from complex compo-
nents. What’s important, rather, is that each component affects something
less than the whole, and so we can partition out the effects of components
more easily.
Fourth and finally, our mechanistic explanations are often indifferent to the
composition of the mechanistic components. Many mechanistic explanations
care only about the location and characteristic activities of the mechanistic
components. Mechanistic explanations often involve functional decomposi-
tions (Cummins, 1980, 1983), and functional decomposition cares only about
the activities of the components and their relation to one another.
Note that this is a weaker claim than saying that mechanistic components
are multiply realizable, at least in the classic sense. Mechanistic components
might well be constrained, and rather severely, by the environment in which
they must work (Shapiro, 2005). I’m inclined to think that myself. The
claim is rather that for many mechanistic explanations, we don’t care about
how the components are composed: as far as the explanation is concerned,
the fuel rail is just something that manages to get fuel from the pump to the
injectors. The details aren’t that important, at least when I’m explaining
hard starts.
2.3 The Hegemony Thesis
The features common to mechanistic components illuminate some of the
advantages of mechanistic explanation. Mechanistic components are per-
sisting, repeatable aspects of causal chains that make a unique contribution
to the behavior of the whole. Philosophers may differ on why things with
those features are especially important features of good explanations—but
few doubt that the do make for good explanations.
I will return to this point further in section 5. First, I want to point out
a presupposition of accounts like Craver’s—one that, if not stated outright,
is strongly implied by things mechanists say. Let’s use “mechanism” to
refer to the whole whose behavior we intend to explain, and “mechanistic
components” to refer to the spatiotemporal parts of that mechanism with the
features just identified. The mechanists’ claim, as I understand it, is that all
good explanation of mechanisms can be cashed out in terms of the activities
5

of the mechanistic components. Mechanistic components might not be the
only things that contribute causally to the explanation of mechanisms. But
they are (in ordinary circumstances) the only things we really care about
when giving explanations of mechanisms.
So for example, mere background conditions also contribute causally to
mechanisms. Engines probably won’t work in the absence of gravity. Never-
theless, good explanations of engines can get by without mentioning gravity.
Call this thesis – that appeal to mechanistic components does the sole and
primary explanatory job in mechanistic explanations — the hegemony thesis.
The hegemony thesis explains why we ought to look for mechanistic ex-
planations, rather than any old causal ones. Mechanistic explanations cite
mechanistic components, which have a variety of attractive features. The
hegemony thesis thus explains why mechanisms are so important.
Or it would if it were true. Alas, there are good mechanistic explanations
of mechanisms that don’t primarily cite mechanistic components.2 It is to
one species of those that I now turn.
3 Resources & Explanation
3.1 Some Examples
My engine stops running. Why? Because the car ran out of gasoline. With-
out gasoline, the engine doesn’t go. That seems, intuitively, like a mechanis-
tic explanation—certainly, it’s on a par with the explanation of my engine’s
failure where I cited a bad check valve. (Many troubleshooting procedures
for engine failure begin by checking whether you ran out of gasoline, be-
fore moving on to components that may have failed). Furthermore, gasoline
seems to pass the constitutive relevance condition: by changing how much
gas there is in the car I can change whether the engine goes, and running
the engine changes how much gas there is.
Yet gasoline is very much unlike the rest of the engine. Gasoline is trans-
2Or, if you want to reserve “mechanistic explanation” for things that do primarily cite
mechanistic components, that there are good explanations of mechanisms that look a lot
like mechanistic explanations except that they don’t cite mechanistic components. I find
the former way—e.g. using “mechanistic explanations” to refer to the broader type—
slightly less clunky. But this is a fight over an explanatory practice, and emphatically not
over the term itself.
6

formed by the action of the engine. The engine persists through those trans-
formations. Gasoline thus gets used by engines, while engines do the using.
To give it a name, gasoline is a resource. There are many resources that show
up in explanations, across a variety of fields. The grass in the paddocks and
water in the streams are resources for the livestock. The catalytic enzyme
argininosuccinate synthase (ASS) is the rate-limiting step in the synthesis
of arginine. Gas, grass, and ASS are thus resources that get used by the
mechanisms of cars, cows, and mitochondria, respectively. They’re also
necessary: you can’t ride for free.
Sometimes resources are familiar and concrete, as in the above. Other re-
sources are more abstract. City planners must consider tax bases, storm
drain capacity, and their ability to provide vital services. Each are re-
sources that a city needs to run efficiently, but each is spread out over
time and space. Computational complexity affects the amount of memory
and processing time that an algorithm needs to run, and is in turn affected
by the number of processors available to implement the algorithm. The
performance of an internet connection depends on resources like bandwidth
and downstream cache. Cognitive scientists appeal to a variety of resources:
working memory capacity, attention, and willpower. We explain character-
istic patterns of psychological performance and deficit by showing how these
resources can be blocked, competed for, and otherwise made unavailable.
3.2 Features of Resources
Examples are easily multiplied (and more will come). I have already pre-
sented a diverse bunch. What do they have in common? I will suggest a
cluster of features, each of which complements those of mechanistic compo-
nents. As with mechanistic components, these are not meant to be necessary
and sufficient conditions — that is not the game, and I’ll explain later why
I think such a search would be fruitless. For now, we’ll just take in some
ways in which resources are similar to each other.
First, tokens of resources rarely persist across the timespan of explanations
that feature them. Cows eat grass. Engines burn gasoline. These activities
irreversibly transform token chunks of resource, and that transformation is
necessary for the proper functioning of the mechanism.
More broadly, resources can be made available or unavailable for transfor-
mation. The availability of resources is often a key feature in explanations.
Working memory has a certain capacity, and holding items in working mem-
7

ory diminishes that capacity. Streaming a movie makes it harder to down-
load system updates, because each takes a certain amount of bandwidth.
Second, and along the same lines, token chunks of resources are not individ-
ually important. That is, it doesn’t (typically) matter which bit of resource
gets used, so much as there’s enough of it to go around. The car factory
doesn’t care which bit of steel gets used: it needs some ingot or other, but
the particular one doesn’t matter. Similarly it does not matter (for most
applications) whether you use the first sector on the hard drive or the ten
thousandth.
Mechanistic parts are individually important precisely because the tokens do
such different things. Resources, by contrast, admit of equipotent divisions.
Every bit of gas is basically as good as every other. These divisions might be
effectively continuous (as with water supplies) or they might come naturally
chunked (as with working memory). Each mechanistic part of the computer,
by contrast, does something individually important—you can’t interchange
the hard drive for the memory.
Third, many resources are causally promiscuous, and perhaps all are intrin-
sically so. That is, resources can interact with many different components of
a complex system, and many different components have an interest in using
that resource. All of the grass in the paddock is available to any of the
cows that want it. Even relatively restricted resources—like, say, modality-
specific buffers—are available to any process that might use that resource
type.
Causal promiscuity is worth noting because many resource explanations in-
volve competition for a limited amount of resource. Conversely, many com-
plex mechanisms have portions devoted to controlling access to resources.
Multitasking operating systems spend a great deal of time mediating con-
flicts over access to memory and processor time. The brain promotes blood
flow to areas that are working harder than usual (Nair, 2005). Even when
resources aren’t actually promiscuous, then, it is often because of deliberate
strategies taken by mechanisms to control access, rather than the intrinsic
properties of the resources themselves.
Fourth, we typically do care about the composition of resources, and oth-
erwise similar resources can’t be substituted willy-nilly. Car engines need
gasoline – it would be nice if something else worked, but nothing works quite
as well. Food chains are vulnerable to collapse precisely because organisms
need specific resources. Ion channels might work in a variety of ways, but it’s
important that they let in only one specific sort of ion. Unlike mechanisms,
8

then, we often do care about the composition of a resource.
Further, even substitutions that maintain functional similarity in a broad
sense can cause a variety of problems. Ethanol-augmented gasoline in your
old car, or a substandard batch of steel in your factory, may not shut things
down — but that substitution is what you’ll have to cite when you want to
appeal why things break, fail, and otherwise fall short of expectations.
The contrasting features of resources and mechanistic components are sum-
marized in table 1
Mechanistic components
Persisting
Resources
Non-persisting
Equipotent
Individually Important
Casually Conservative
Realization-irrelevant
Causally Promiscuous
Realization-relevant
Table 1: Contrasting components and Resources
Each of these contrasts should be read as relative rather than absolute.
Sometimes we care about what mechanistic components are composed of,
and sometimes the composition of a resource is irrelevant. Sometimes parts
of a mechanism are destroyed in the course of action. On the whole, how-
ever, there are systematic differences between mechanistic components and
resources.
3.3 Resource Explanations
Resources appear to be the primary explanans in a variety of mechanistic
explanations. There are explanations that account for the activity of the
whole in more or less the same way that canonical mechanistic explanations
work, but which appeal to resources rather than mechanistic components.
Again, examples will make the point.
Most obviously, many explanations must cite resources if they’re to be com-
plete. You don’t know how a car factory works unless you know about the
materials it uses and the machines that do the using. An explanation of
how a heart works is obviously incomplete without mentioning the blood
that gets pumped around. In many such explanations, resources might not
play the most interesting causal-explanatory role, but they certainly play
9

an ineliminable one.
Further, many explanations give resources pride of place. Some of these
invoke various ways in which resources can be depleted, blocked, competed
for, or otherwise made unavailable. Availability of resources thus shapes
the behavior of the whole. Sometimes this relationship is crude: no gas,
no driving. But the relationship can be more complex. Forming long-term
memories via long-term potentiation uses glucose, which is largely provided
by astrocytic glycogen. Depleting glycogen experimentally leads to amne-
sia. The elderly often have memory problems, and also tend to have lower
levels of glucose. Supplementing them with glucose improves their memory
(Messier, 2004; Newman et al., 2011). Crucially, the relationship between
glucose and LTP is not a simple on-off sort of relationship: the relationship
is a specific one, in the sense specified by (Woodward, 2010).
Computational complexity theory similarly posits specific relationships be-
tween an algorithm’s properties and the amount of time, space, and so forth
it can use. Consider: many low-level assembly language algorithms are tuned
to avoid accessing RAM, which can be orders of magnitudes slower than ac-
cessing in-CPU cache (Duntemann, 2011). The performance of instances of
the algorithms can be complex, but largely explained by the differences in
access time between two distinct types of memory resource. Along the same
lines, some enzymes are the rate-limiting step in reactions, and so the norm
of reaction within the mechanism is largely explained by reference to the
bulk availability of that enzyme.
Appeal to resources can also be evidentially important, because they serve
to invalidate (or at least cast suspicion upon), certain types of mechanistic
explanation. Tim Shallice’s relies on considerations of resources in his argu-
ment that single dissociations offer only weak evidence for the distinctness
of two cognitive processes (1988). Suppose I have a lesioned patient who
can (say) spell regularly spelled English words and regular nonwords, but
not irregular words. I might conclude that there are two routes to spelling:
a preserved phonetic one that involves spelling-to-sound correspondences
and a damaged lexical one that retrieves stored information about spelling.
But as Shallice points out, there is another possible explanation: spelling
irregular words might just be harder. That is, spelling might have only one
route, but that route might require a general resource that is stressed more
by harder tasks than by easier one – intelligence or working memory or at-
tention, say. Partial damage to this resource will only affect more difficult
tasks. Ruling out these resource artifacts is the primary reason why double
dissociations are so prized in neuropsychology.
10

Note too that this appeal to resources and their properties are being done
within the context of straightforwardly mechanistic box-and-arrow sorts of
cognitive psychology. There’s no need to go further afield to networks or
dynamic systems to press the point. Even on its home turf, mechanistic
explanation appears to leave something out.
Further, many of the techniques for distinguishing and exploring resources
involve the same kinds of scientific methods that picking out mechanistic
parts do. Just as it is not always obvious which spatiotemporal parts count
as mechanistic parts, it is also not obvious which resources a mechanism ac-
tually uses. The very same techniques for exploring mutually manipulability
that Craver details in his (2007) can be applied to resources. So for example,
to determine whether something is a resource you can manipulate it, block
access to it, change the whole to measure the amount of the resource, and
so on. So it seems that scientists do very similar things to investigate both
mechanistic parts and resources. The two thus deserve explanatory parity.
Finally, differences between resources can often explain differences between
mechanisms. Sometimes these involve different resources as inputs. Different
steelmaking processes, for example, vary importantly in what resources they
use (forced air in the Bessemer process vs pure oxygen in contemporary BOS
methods, say). Differences in resources often require different mechanisms:
diesel engines don’t have spark plugs and have heavier combustion chambers,
for example, because diesel must be detonated via high compression rather
than merely ignited. Explaining why the two types of mechanism differ thus
cites the different resources those mechanisms are meant to use, even if they
are functionally similar at the level of the whole.
At a more abstract level, resources themselves can vary in a variety of ways.
As an incomplete taxonomy: resources can be limited or unlimited, they can
admit of discrete or continuous divisions, they can be permanently trans-
formed or merely blocked, usable in parallel or in serial, and so on. These
distinctions make a difference to the systems that use them. Computers
with random-access memory have different properties to older serial-access
kinds. Omnivores have a different ecological profile than carnivores. Break-
bulk shipping has a different timescale and efficiency curve to container
shipping. The differences between these complex mechanisms is explained
by the differences in the resources they use. Resource explanations are thus
important when we compare mechanisms as well as when we try to account
for the behavior of single mechanisms.
In sum, resources constitute their own class of explanans, distinct from mech-
anistic components, and sometimes forming the primary or even the sole ex-
11

planatory variable for the behavior of mechanisms. The hegemony thesis is
false: explaining mechanisms can involve more than discussing mechanistic
components.
4 Interlude: Objections and Replies
4.1 Why Not Just be Broader?
The hegemony thesis states a strong restriction on the resources available to
mechanistic explanation. One might object that it is unreasonably strong.
Mechanists, the objection goes, have always included resources in their ex-
planatory arsenal.3 After all, Craver’s mutual manipulability condition picks
up resources too. William Bechtel has been especially clear about this, writ-
ing
. . . a part consists of components and operations. By parts I
designate the structural components of a mechanism whereas by
operations I refer to processes or changes involving the parts.
I use the term operation to emphasize that in each operation
not only is something performing an operation but something
is operated on. In chemistry, investigators could identify the
chemical substances (substrates) that are changed in a given
reaction, and use that in characterizing the reaction. In the
cognitive domain, investigators speak of information processing,
suggesting that information is what is operated on. . . (Bechtel,
2008, 14)
Bechtel is clear: there are simply parts of a mechanism, and some of them
are operated upon while others do the operating. Resources fall into the
former category. Mechanistic explanation was never incomplete.
My response has several aspects. As far as sociology of philosophy of science
goes, I submit that the attention of the mechanists has overwhelmingly been
focused on mechanistic components. Even if some definitions managed to
include resources, discussions of mechanistic explanations themselves almost
always focus on the components doing the manipulating.
3Thanks to Elizabeth Schier for pressing me on this objection.
12

Second, as I’ve shown, resources have very different properties to mechanis-
tic components. Thus, even if they are both (broadly construed) parts of
mechanisms, there are good scientific reasons for treating them as distinct
kinds of things.
Third, resources are not always spatiotemporal parts of mechanisms. Many
resources — the grass in the paddocks, the water in the river, the bandwidth
of the wifi network — are external to the mechanisms they help explain.
They are not located within the mechanism, and it is unreasonable to extend
the mechanism to include them. The boundary between mechanisms and the
resources they use can be contentious—drawing that boundary in the case
of cognition is at the center of the extended mind debate (Rupert, 2009).
This is a further reason, then, to distinguish mechanisms and resources.
Finally, there is a good philosophical reason to distinguish mechanistic com-
ponents. In section 1, I noted that a persistent problem for mechanists was
explaining why mechanisms — as opposed to mere causes — were explana-
torily important. The picture I sketched of mechanistic components gives
one answer: mechanistic components tend to persist through a variety of
different causal processes. Mechanistic components provide fixed points in
the causal flux that can be related together to give an overall picture of the
whole. Further, the same components can appear in a variety of different
explanations, unifying them under a common ontological framework.
Resources, since they don’t reliably persist, can’t play the same justificatory
role. On the other hand, they can play a different, distinct, and equally
valuable role in explaining why mechanisms are the way they are. The
realization-dependence of resources, for example, explains and thus anchors
the form of mechanistic components. There are many ways one might build
a check valve, but if it’s a check valve for high pressure gasoline the available
options are greatly reduced. This sort of information potentially gives depth
and unity to otherwise disconnected causal explanations. Resources also
vary much more frequently and variously than mechanistic components –
in their presence, in their levels, in their composition, and so on. These
variations are actual difference-makers across a wide variety of mechanisms,
and so worthy of our attention.
In sum even if mechanists haven’t distinguished components and resources,
they have good reason to do so.
13

4.2 But aren’t mechanistic components more basic?
A second objection notes the apparent passivity of resources. Mechanistic
components sometimes work on their own. Clocks, those paradigms of mech-
anistic virtue, can be explained by reference only to their mechanistic com-
ponents. Resources need mechanisms, on the other hand. Since resources
are used, they need something to do the using. Mechanistic components are
thus the more basic explanatory notion.
This objection must be phrased with care. For the issue is not whether
resources are ever the only things appealed to in an explanation—for the
sake of argument, let us grant that they aren’t. Resources can still be the
most important aspect of the explanation by being—for example—the only
things that vary, and so the only actual difference-maker in a population of
otherwise similar mechanisms.
The point could be read as a meta-philosophical one rather than one about
particular explanations. Mechanistic components must be conceptually or
logically or explanatorily prior to resources. We only care about resources,
this line goes, because we care about the mechanistic components that use
them.
This version also seems implausible to me. While some mechanistic com-
ponents can be understood and appealed to without mentioning resources,
many of them can’t be. An explanation of how an engine worked that didn’t
mention gasoline would be hopelessly incomplete. Many mechanistic com-
ponents can only be understood by understanding the sorts of resources
they’re meant to use: a fuel injector is something that squirts gasoline, and
if you miss that you’ve missed the point. Even if not all mechanistic expla-
nations involve resources, this wouldn’t entail that resources aren’t a vital
and equal feature of the explanations in which they do appear.
A final way to run the objection appeals to different levels of explanation.
So far, I’ve spoken of things as if they are either mechanistic components or
else resources, and that this categorization can be made in some objective,
atemporal sense. That is not so. The hierarchical decomposition strategy of
mechanistic explanation often means that what is a mechanistic component
in one explanatory context can be a mechanisms in another context. The
capacities of the fuel rail can themselves be explained by treating it as a
mechanism in its own right, and looking to its on mechanistic components.
Resources can also be treated as mechanisms. RAM is a resource for the pro-
grammer, but a collection of mechanistic components for the chip designer.
14

Computational neuroscience seeks to undercover the mechanisms that un-
derly attention. Many purported resources, then, become mechanisms upon
closer look. Perhaps that shows that mechanistic components are the more
basic notion?
Alas, hierarchy cuts both ways. Mechanistic components can also be re-
sources in some contexts. Kidneys are mechanistic components for the
anatomist, a resource for the butcher. Jeeps are complex mechanisms to
the motor pool, mat´eriel to the general. Nor is there a general principle
that the lower level is always the mechanistic one. RNA is a mechanism
when you care about transcription, a cellular resource when you care about
viral load. Both of those are explanations situated on roughly the same
cellular-biological level. Water towers are (simple) mechanisms; part of the
explanation of how they work treats the same object as including a very large
spatial resource. More generally, the fact that some explanations of mecha-
nisms appeal to resources (established in section 3.3) shows that sometimes
resources can be lower-level explanatory aspects of a mechanistic component.
Mechanistic explanation, then, does not privilege mechanistic components
as ontologically more basic or otherwise special. Mechanistic components
and resources are equal partners in explanatory projects.
4.3 Aren’t resources background conditions?
A final objection (and possibly the most pressing) seeks to assimilate re-
sources to background conditions. On this line, resources do figure in mech-
anistic explanations. However, they do so only in a secondary or derivative
sense, by providing the necessary background canvas upon which the real
explanation can be sketched. Of course (the line goes) engines need gasoline
to run, and wouldn’t without. But they probably also wouldn’t run in a
vacuum, or in zero gravity, or in the presence of serious amounts of ioniz-
ing radiation. All of these specify background conditions that have to be in
place before mechanistic explanations can get off the ground. But given that
they are in place they – well, fade into the background. At best, they are
appealed to to explain why mechanisms fail to work, not why they succeed.
That’s always a derivative project, though: failure is intelligible only against
a background of success.
I suspect that something like this thought has occurred to many, and ex-
plains why resources have received comparatively less attention. I think the
objection is misguided, thought. For starters, there is something more than
a little odd about saying that you understand (say) how a factory works
15

without understanding how it transforms the raw materials that come in.
After all, that is the whole point of the factory. Car factories aren’t com-
plicated clockworks existing for their own sake, merely relying on a supply
of steel to keep things moving. Every mechanistic component is geared to-
wards transforming steel into cars, and you’re surely missing something if
you treat the presence of steel as a mere background condition. (Indeed,
it seems to me more plausible to say that the presence of a factory is the
background condition necessary for your factory to keep turning boring steel
into shiny cars).
More technically, the standard marks of ‘mere’ background conditions don’t
seem to apply to many resources. First, resources can’t be just bracketed
off in mechanistic explanations. Some are, in Kenneth Waters’ term, ac-
tual difference-makers (2007). That is, the level of some resource actually
varies in populations we care about, and that actual variation is part of the
explanation of actual differences between mechanisms in that population.
To return to a few cases: the same computational mechanism might vary in
efficiency depending on how much memory it has available. The mechanisms
underlying obesity are sensitive to differing proportions of carbohydrates
and protein in food. Breathing rate and depth are determined by the level
of carbon dioxide in the blood. Urea synthesis rates depend on relative
availability of various enzymes. And so on and on. Note that all of these
are explanations about facts of normal functioning. It’s true that, in each
case, the lack of the resource would cause the mechanism to fail. But that
can hardly be a condition for explanatory irrelevance: pull a spark plug out
of your engine and you’ll get equally dramatic failure.
That leads directly to the second point. In a well-known piece, James Wood-
ward notes that different causes might have more or less specific relationships
to their effects (Woodward, 2010). Roughly speaking unspecific causes have
few possible states that lead to few possible states of the outcome (at the
limit case, just two states with a ‘switch-like’ relationship to the effect),
while specific causes have many states that stand in a more or less one-to-
one relationship to many distinct effect states. Woodward notes that when
X is a relatively non-specific cause of Y , then “we are more likely to regard
X as a mere enabling (or background) condition for Y” (Woodward, 2010,
317). Craver suggests a slightly different criterion for background condi-
tions. A background condition, he notes, typically won’t pass the mutual
manipulability test. I might be able to affect my engine’s functioning by
changing the local gravitational field, but I can’t change gravity by running
the engine Craver (2007).
16

Both of these criteria seem to me like good rules of thumb for picking out
mere background conditions. Many resources don’t satisfy either of them.
To recall one example: astrocytic glycogen is a crucial energy source for
long-term memory formation. Forming memories depletes astrocytic glyco-
gen, and depriving astrocytes of glycogen causes amnesia. So astrocytic
glycogen satisfies Craver’s mutual manipulability criterion. Further, there
is a sensitive relationship between glycogen and performance on memory
tasks Newman et al. (2011): a little depletion causes a little deficit, more
causes more, and so on. So astrocytic glycogen passes Woodward’s test as
well.
Of course, not all resources will pass both tests (though I conjecture that
most will pass at least one). There is a further, trickier question about how
to deal with resources that do not coincide the spatiotemporal boundaries
of the mechanisms (as with, for example, the ambient air), though I suspect
those should also count as resources at least in some cases.
Despite these complications it’s obvious that there’s a deep explanatory dif-
ference between resources and mere background conditions. Indeed, the
point can be turned around. I suspect that some of the difficulty in coming
up with an adequate characterization of background conditions (at least for
mechanistic explanations) might stem from the fact that mere background
conditions are not reliably distinguished from resources. All the more rea-
son, then, for philosophers to distinguish the two.
5 Mechanistic Explanation Reconsidered
To conclude, I’d like to return to the question that I raised at in section 1.
Given that scientists in many disciplines apparently do care about mechanis-
tic explanation, why is that the case? That is, why care about mechanisms,
rather than just about causes?
I gestured at one possible answer in section 2.2. Mechanistic components
tend to persist over time, and so explanations involving them have a certain
kind of depth or unificatory power that would be lacking in explanations
that cite ‘mere’ causes (whatever those would be). Still, one might find
this less than completely satisfying. After all, causal explanations are also
enhanced by describing stable features of the causal network4
4Just why that’s an improvement is an open question. Most mechanists appeal to
higher-order properties. i’m a fan of theories on which the benefits are pragmatic rather
than metaphysical in nature; see my Klein (2014b) for some arguments.
17

We can say more. If it was merely a matter of persistence over time (say),
resources wouldn’t have the explanatory status that I’ve argued they actu-
ally have. Rather, I think the mechanist project is actually broader still.
The goal of mechanistic explanation is to taxonomize causes (implicitly or
explicitly), clustering them into groups with similar features. I’ve men-
tioned three such clusters so far: mechanistic components, resources, and
background conditions.
Mechanistic components and resources, on the other hand, are distinguished
by a variety of more fine-grained features. Again, these features are explanation-
relative: a resource in cognitive science might be treated a mechanism in
neuroscience. The carving is not arbitrary, however, and we can identify
features of explanations that determine where an object falls in the taxon-
omy.
One obvious candidate for a taxonomic principle is the timescale of expla-
nations. Processes that are effectively stable at one timescale can fluctuate
at larger ones; conversely, fluctuating processes can exhibit longer-term sta-
bility at a coarser grain. Mental processes occur at a variety of timescales,
from milliseconds to years (Newell, 1990, 80-81). One goal for mechanistic
explanations is mechanisms is to find things that persist for some useful
length of time, but “useful length of time” is obviously going to depend in
part on the explanatory project.
Another candidate is the specific functional relationships in which an en-
tity stands. RNA, for example, might count as a resource in explanations
that treat it as an input to protein synthesis, but a mechanistic component
when we treat (say) DNA as the input and RNA as a component in the
mechanism.5
Put this way, there’s nothing special about mechanistic components per se.
Mechanistic components are explanatorily important precisely because they
share features that make for good explanations. Resources similarly share
distinct, but important, features, and these also make them good candi-
dates for explanation. Their ongoing change and flux, for example, makes
them excellent candidates for the actual difference-makers in a system. Fur-
ther, even when they’re stable, resources have a different modal profile than
mechanistic components, and so support different kinds of explanations.
Perhaps the most abstract distinction, however, is that between the agent
and the patient of action. Mechanistic components are agents: they do
things. They might be caused to do things, but they are primarily active.
5Thanks to Chris Lean for this suggestion.
18

Resources, by contrast, are largely patients. They are acted upon. We speak
about resources very naturally in the passive voice (“Gasoline was burned.”)
Their passivity is why we need mechanisms to control their availability and
influence (Money doesn’t do anything without people to move it around,
banks to keep it safe, and so on). “Agent” and “patient” are, again, relative
to the kinds of activity in question, which in turn leads to the explanation-
relativity of the categories themselves.
Thinking of it this way makes clear that the categories of mechanistic com-
ponent and resource might themselves be neither necessary nor exclusive.
For disciplines without a clear agent-patient distinction, the two carvings
might not get much purchase. Simple physical explanations are an obvious
example: in the world of atoms and void, things simply happen, without
actor or acted upon. This is why the mechanistic style of explanation gets
little purchase in basic physics.
This is not something special about physics, though. Even more complex
physical explanations might reproduce the agent-patient structure. Consider
explanations of supernovas, for example: the star uses its fuel, and once it
is forced to rely on inferior resources for stellar combustion, it goes nova.
Having reconstructed the agent-patient structure, it also becomes natural
to talk about the mechanisms of stellar evolution.
Conversely, there are systems that resist an easy agent-patient dichotomy
because they appear to have features of both at once. The most obvious of
these are what Bechtel calls active mechanisms (2008, Ch8). These include
the mechanisms involved in autocatalytic loops, homeostatic processes, and
other feedback cycles. Key to mechanisms with feedback is the possibility
of a mechanism altering its own operation — that is, acting at the same
time as agent and patient. Active mechanisms are, as Bechtel argues, key to
understanding many living processes; they’re also arguably more important
for understanding many neural processes (Bechtel, 2008; Klein, 2014a).
Where does this leave us? I think we might conclude with a complex but
realistic pluralism. Cataloging the things that go into mechanistic explana-
tions — and, indeed, mechanisms, and cataloging things as mechanisms in
the first place — is relative both to explanatory projects and how the world
is structured. If that’s true then we can see why mechanistic explanation is
important but not fundamental. It is a strategy that some fields will employ
more than others in the course of coming up with good causal explanations.
But that is not to downplay the importance of the mechanistic project.
The details of mechanistic explanations will be heterogenous, and differ-
19

ent projects may even come up with different ways of carving the world.
Most obviously, disciplines which frequently encounter active mechanisms
will have a more complicated time distinguishing pure mechanistic compo-
nents, mere resources, and things that have features of both. Nevertheless,
we have an answer to why we prefer mechanistic explanations when they are
available, rather than just a causal/interventionist story. They pick out cat-
egories of explanans that share certain broad features that make for good
explanations. I’ve suggested that there are more distinct categories than
typically supposed, and that these deserve respect as well. Resources are an
additional category that deserves special respect.6
References
Bechtel, W. (2008). Mental mechanisms: Philosophical perspectives on cog-
nitive neuroscience. Taylor & Francis, New York.
Bechtel, W. and Richardson, R. (2010). Discovering complexity: Decom-
position and localization as strategies in scientific research. MIT Press,
Cambridge, mit press edition. edition.
Craver, C. (2007). Explaining the brain. Oxford University Press, USA.
Craver, C. F. and Darden, L. (2013). In search of mechanisms: Discoveries
across the life sciences. University of Chicago Press, Chicago.
Cummins, R. (1980). Functional analysis. In Block, N., editor, Readings in
the Philosophy of Psychology, volume 1, pages 185–191. Harvard Univer-
sity Press, Cambridge, MA.
Cummins, R. (1983). The nature of psychological explanation. MIT Press,
Cambridge.
Duntemann, J. (2011). Assembly language step-by-step: Programming with
Linux. John Wiley & Sons.
Kaplan, D. M. and Craver, C. F. (2011). The explanatory force of dynamical
and mathematical models in neuroscience: A mechanistic perspective.
Philosophy of Science, 78(4):601–627.
Klein, C. (2012). Review of Explaining the Brain, by Carl F. Craver. Mind,
121(481):165–169.
6Thanks to Liz Irvine, Chris Lean, Elizabeth Schier, and participants at the 2014
Cognitive Science Kioloa workshop for helpful comments on an earlier draft.
20

Klein, C. (2014a). The brain at rest: What it is doing and why that matters.
Philosophy of Science, 81:974–985.
Klein, C. (2014b). Psychological explanation, ontological commitment, and
the semantic view of theories. In Sprevak, M. and Kallestrup, J., editors,
New Waves in Philosophy of Mind, pages 208–225. Palgrave Macmillan,
New York.
Levy, A. and Bechtel, W. (2013). Abstraction and the organization of mech-
anisms. Philosophy of science, 80(2):241–261.
Machamer, P., Darden, L., and Craver, C. (2000). Thinking about mecha-
nisms. Philosophy of Science, 67:1–23.
Messier, C. (2004). Glucose improvement of memory: a review. European
journal of pharmacology, 490(1):33–57.
Nair, D. G. (2005). About being BOLD. Brain Research Reviews, 50:229–
243.
Newell, A. (1990). Unified Theories of Cognition. Harvard University Press,
Cambridge.
Newman, L. A., Korol, D. L., and Gold, P. E. (2011). Lactate produced
by glycogenolysis in astrocytes regulates memory processing. PloS One,
6(12):1–11.
Piccinini, G. and Craver, C. (2011). Integrating psychology and neu-
Synthese,
roscience: Functional analyses as mechanism sketches.
183(3):283–311.
Rupert, R. (2009). Cognitive Systems and the Extended Mind. Oxford Univ
Prress, New York.
Shallice, T. (1988). From Neuropsychology to Mental Structure. Cambridge
University Press, Cambridge.
Shapiro, L. (2005). The mind incarnate. MIT Press, Cambridge.
Simon, H. A. (1996). The Sciences of the Artificial. MIT Press, Cambridge,
3rd edition.
Stepp, N., Chemero, A., and Turvey, M. T. (2011). Philosophy for the rest
of cognitive science. Topics in Cognitive Science, 3(2):425–437.
Waters, C. K. (2007). Causes that make a difference. The Journal of Phi-
losophy, 104(11):551–579.
Woodward, J. (2010). Causation in biology: Stability, specificity, and the
choice of levels of explanation. Biology & Philosophy, 25(3):287–318.
21