HOW FUNCTIONAL DIFFERENTIATION

ORIGINATED IN PREBIOTIC EVOLUTION

ARGYRIS ARNELLOS
ÁLVARO MORENO

ABSTRACT. Even the simplest cell exhibits a high degree of functional differen-
tiation (FD) realized through several mechanisms and devices contributing
differently to its maintenance. Searching for the origin of FD, we briefly argue
that the emergence of the respective organizational complexity cannot be the
result of either natural selection (NS) or solely of the dynamics of simple
self-maintaining (SM) systems. Accordingly, a highly gradual and cumulative
process should have been necessary for the transition from either simple
self-assembled or self-maintaining systems of functionless structural compo-
nents to systems with FD. We follow results of recent in vitro experiments with
respect to competition among protocells, where a primitive type of selection
begins to operate among them accompanied by a parallel evolution of their
functional domain. We argue that minimal forms of FD should be established
within the evolution of SM processes in protocells as they undergo a simpler
selection process for stability and persistence in a prebiotic environment. We
then suggest the concept of closure of constraints (CoC) as a way to identify and
describe minimal FD in a far-from-equilibrium SM organization. We show in
detail how the concept of CoC together with the conditions for its fulfillment
can be applied in the case of a simple protocellular system that begins to couple
internal chemical reactions with the formation of its membrane components.
Finally, we discuss how such SM systems can evolve towards significantly
higher levels of FD, suggesting this is mainly the result of functional recombi-
nation (formation of mechanisms) in the context of a modular SM organization.
KEY WORDS. Functional differentiation, self-maintenance, natural selection,
constraints, closure, prebiotic evolution, organizational complexity, protocells.

INTRODUCTION
Biological systems are highly complex organizations consisting in many
different sophisticated molecules (e.g., DNA, RNA and proteins) and other
supra-molecular structures (ranging from, e.g., lipid and phospholipid
membranes to tissues, organs, etc.). Those molecules and structural aggre-
gates either based on their properties or, perhaps more importantly, on
the organizational properties emerging out of the specific and intricate
relationships they establish and manage to maintain among them, are

IAS-Research Centre for Life, Mind and Society, Department of Logic and Philosophy of
Science, University of the Basque Country, Donostia - San Sebastian, Spain.
argyris.arnellos@ehu.es / alvaro.moreno@ehu.es

Ludus Vitalis, vol. XX, num. 37, 2012, pp. 1-23.

2 / LUDUS VITALIS / vol. XX / num. 37 / 2012

considered as playing different functional roles in the dynamic behavior

of the organizations in which they participate. As such, biological systems
and, actually, even the simplest cell, exhibit a high degree of functional
differentiation (FD) realized through mechanisms and devices that ensure
their survival and adaptation.
However, it is sensible to think that even the simplest present-day cell
organization is much more complex than what would be required for
harboring a minimal form of FD. So, where did this FD come from and how
was it developed? Several approaches resort to the mechanism of natural
selection (NS) to ground the concept of function (Millikan 1989; Neander
1991; Godfrey-Smith 1994). But as it explained in Fox Keller (2007, 2010)
and in Moreno and Ruiz-Mirazo (2009), the way in which NS is conceived
within Darwinian theory requires an open, sufficiently wide phenotypic
domain where selective forces may act. In turn, underlying the idea of
phenotypic domain one can find an implicit, usually not very well defined
idea of ‘functional domain,’ which refers to the variety of ways in which
the different (structural) components of a system interact and complement
with each other in their continuous contribution to the global maintenance
of that system. This functional domain is already implicit in a minimal, but
sufficiently rich concept of a cell.
This was not the case in prebiotic evolution. It is widely accepted that
the organizational complexity and the respective functional domain of the
simplest present-day cells should have been preceded by several infrabi-
ological forms of organization. As it is argued in Moreno and Ruiz-Mirazo
(2009), there are several experimental cases (e.g., Walde, et al. 1994; Segre
and Lancet 2000, and Chen, et al. 2004), which have shown that multipli-
cation, variation and heredity would constitute a necessary but not suffi-
cient set of conditions for what is commonly understood as evolution by
NS. There was for sure a time when the types of evolution were not based
on the Darwinian NS, when NS was not fully operating (Godfrey Smith
2009).
Considering that there were systems that fulfill those conditions but do
not evolve as present-day biological systems do, it is sensible to think that
NS is the result of an already evolved functionality. As it shall be developed
in section 3 (and also as has been argued in Moreno and Ruiz-Mirazo 2009),
NS starts with a system able to exhibit some sort of structural as well as
functional variability (i.e., an already functionally differentiated system).
Therefore, NS cannot account for the emergence of minimal FD; on the
contrary, NS needs FD in order to be operative. Accordingly, we have to
conclude that a more or less long cumulative process has been necessary
for the emergence of FD in the cells.
The structure of this paper is as follows: In section 1 we argue that a
certain type of far-from-equilibrium self-maintaining (FFE SM) systems,
 ARNELLOS; MORENO / FUNCTIONAL DIFFERENTIATION / 3

based on chemical autocatalysis associated with self-assembly, seem to

have some properties that can support the gradual process needed for the
emergence of FD. Then, we argue that the concept of function before the
appearance of NS should be considered in conditions that favor a simpler
selection process for stability and persistence of self-maintenance organi-
zations. In section 2, based on in vitro experiments, together with relevant
in silico results, we suggest that the organizational complexity necessary
for minimal FD emerges in cases where simple protocells begin to couple
internal chemical reactions with the formation of their membrane compo-
nents, in a cascading of primitive competitive-evolutionary selection dy-
namics. In section 3, we introduce the notion of closure of constraints (CoC)
in order to present a detailed account of the conditions and requirements
for minimal FD, and to provide a way to identify and describe minimal FD
in a system. In section 4, we discuss how FFE SM systems with FD can evolve
towards significantly higher levels of FD in an open way.

1. FUNCTION PRIOR TO NATURAL SELECTION

1.1. SIMPLE SELF-MAINTAINING SYSTEMS

ARE NOT ENOUGH FOR MINIMAL F.D.
The main scientific alternative to evolution through NS for the explanation
of the evolution of FD, are the dynamics of a self-maintaining (SM) system
emerging out of self-organizing processes. Self-maintenance is a wide-
spread phenomenon exhibiting different levels of organizational com-
plexity, but its minimal expression consists of the so-called ‘dissipative
structures’ (Glansdorff and Prigogine 1971; Nicolis and Prigogine 1977;
Nicolis 1989). These are forms of self-organization that result in the produc-
tion of systems that are maintained in a steady state through a flow of
available energy. In such systems and in the presence of a specific flow
of energy and matter in far-from-equilibrium (FFE) thermodynamic con-
ditions, a huge number of microscopic elements adopt a global, macro-
scopic ordered pattern (a ‘structure’), which in turn contributes to the
maintenance of the FFE flow of energy and matter enabling the persistence
of the microscopic dynamics.
The interest of the concept of self-maintenance is that it relates function
to the self-maintaining system by interpreting it as a specific causal effect
of a component or trait which contributes to the maintenance of the system
and, consequently, of the component itself in an organizationally closed
way. Then, in this context appears a basic idea of functionality understood
as the contribution of a component to the self-maintenance of the system
(Schlosser 1998; Collier 1999, 2000, 2004; Bickhard 2000, 2004; McLaughlin
2001; Christensen and Bickhard 2002; Delancey 2006; Edin 2008 and Mos-
4 / LUDUS VITALIS / vol. XX / num. 37 / 2012

sio, et al. 2009). However, is the concept of simple self-maintenance

enough for an explanation of the emergence of FD in biological systems?
Let us consider a candle flame as an example of a relatively simple SM
system. In a candle flame the microscopic reactions of combustion gener-
ate a macroscopic pattern (the flame itself) that contributes to maintain the
conditions for its own existence. Specifically, the flame makes a crucial
contribution to maintaining the microscopic chemical reactions by keep-
ing the temperature above the combustion threshold, vaporizing wax, and
inducing convection (which pulls in oxygen and removes combustion
products). In turn, the chemical reactions keep generating the combustion
and thus the flame itself (Bickhard 2000). In particular, the emerging
configuration (i.e., the flame) constrains the surroundings (temperature,
wax, oxygen) by turning them into appropriate boundary conditions
required for its own maintenance. As mentioned elsewhere (Mossio and
Moreno 2010), the candle flame is a simple SM system, whose maintenance
depends on the work of a single constraint acting on its own boundary conditions
(see figure 1). The candle flame works to maintain the boundary condi-
tions that enable it to keep on working. In all, a candle flame can do only
one thing, which is equal to saying that a candle flame has only one
function. Indeed, a candle flame has no options to select between alterna-
tive ways to enable its continuous maintenance, and this result in its
incapability to correct the threat of getting extinct when the boundary
conditions are not the appropriate ones (i.e., running out of wax, decrease
of oxygen, etc.).

FIGURE 1.
Abstract depiction of a simple SM system whose maintenance depends on the
work of a single constraint (C) acting on its own boundary conditions (Bd_C).
It can be the case of a candle flame that constrains its surroundings by turning
them into appropriate boundary conditions for its maintenance.

Anyhow, certain SM systems (as we shall discuss later) can reproduce

themselves and therefore support some form of continuity allowing, at
least in principle, an accumulation of novelties. These systems are poten-
 ARNELLOS; MORENO / FUNCTIONAL DIFFERENTIATION / 5

tially capable of producing and preserving new components with increas-

ing structural complexity. Interestingly, these systems may recruit some
of the new components to act as constraints to modify their own organi-
zation, which is precisely the scenario we are looking for the origin of FD.
However, in order to explain the emergence of a minimal form of FD, it
would be necessary that these systems show differentiated paths in the
contribution to their self-maintenance in such ways that NS can operate.
How then can we explain the primitive cumulative process of organ-
izational complexification up to the level of minimal FD, from chemical
evolution to systems able to evolve by the mechanism of NS?

1.2. FUNCTION PRIOR TO NATURAL SELECTION

THROUGH ‘SELECTION FOR PERSISTENCE’
In several papers analyzing the origin of functions in prebiotic evolution,
Fox Keller (2007, 2010) lays out what she considers as a minimalist concep-
tion of the notion of function. She specifically argues, as we also mentioned
in the introduction, for the need of an already evolved functionality for NS
to be operative. Additionally, she characteristically argues that in this
context of pre-NS competitive-evolutionary dynamics, one should speak,
in terms of persistence—rather than survival—of the system and capacity
for robust self-maintenance—instead of fitness.
Fox Keller suggests that the origin of functionality could be traced down
to systems that pose some simple form of self-regulation. According to her
thesis, once such ‘mechanisms’ are observed, we could legitimately speak
of function in a system. What Fox Keller is trying to stress is that, for a
concept of function that preceded NS, what is needed is not a notion of
“usefulness of an entity to a system” as strong as implied by the conditions
that make NS operative, but rather the existence of a property that contrib-
utes to the persistence of the system to which that property belongs, which
will be enough to trigger a different (simpler) selection process for stability
and persistence. In other words, the application of feedback mechanisms,
which result in the generation of stabilities in a system, is enough for the
operation of simpler evolutionary-competitive dynamics that will result
in different degrees of maintenance to those stabilities. She then ascribes
functionality to these primitive ‘regulatory’ mechanisms, since they con-
tribute to the maintenance of the stabilities.
Further, the capacity for regulation requires the modulation of an
underlying range of functional operations on the structure of each trait,
thus implying a higher degree of organizational complexity than what is
likely to be found in prebiotic systems. Actually, Fox Keller’s account is not
very minimalist and this is evident in the examples she introduces in order
to demonstrate her concept of minimal function. Indeed, the chemotaxis
of E. coli and the regulation of oxygen in a termite’s nest, due to the
6 / LUDUS VITALIS / vol. XX / num. 37 / 2012

dynamic construction of a mount by the termites themselves, may be

examples of several feedback mechanisms added and nested together into
highly complex systems in the course of evolution, where the respective
functionality presupposes a high degree of organizational complexity.
Therefore, we should look for much more simple SM systems that may
show a minimal form of ‘selection’ for persistence.
As we discussed in section 1.1., the minimal expression of an organiza-
tion that persists is realized through a dynamically FFE, self-constitutive
and dissipative kind of system, where a set of randomly interactive struc-
tural components bring forth a global self-maintaining arrangement. The
components of such organization must allow a variety and plasticity of
interactive relations. In a prebiotic environment such organization can
produce new structural components (in an open way, and based on
recombinations), generating this way an increase of different interactive
relations. In addition, some components (catalysts) can selectively modify
the composition ‘rules’ between the other structural components, thus
generating internal structural variety. Provided a long and wide scenario,
where adequate boundary conditions can be fulfilled, this organization
has the capacity to explore and retain a wide variety of this type of
self-(re)producing SM systems.
This brings up an important issue. Under suitable boundary conditions
those systems could persist and produce new structural components. Hence,
what we should look for is how new, more complex components can
contribute to the maintenance and persistence of the organization they
belong to.
Recapitulating, in this section we commented on the implications for
the capability of simple self-(re)producing SM systems to persist in a
prebiotic environment with respect to their functionality. We suggested
that in principle, simple self-(re)producing SM systems might achieve a
robust persistence and evolve towards systems with greater organiza-
tional complexity by recruiting and combining structural complexity for
new and different organizational tasks.
Yet how can prebiotic systems evolve towards increasingly complex SM
systems exhibiting a minimal form of FD? As we will see in section 2, this
can be achieved by combining processes of self-assembly and self-organi-
zation in certain forms of self-reproducing SM systems.

2. EXPERIMENTAL RESULTS AND MODELS

OF SELF-MAINTAINING PERSISTENCE IN PREBIOTIC EVOLUTION
In an attempt to understand the emergence of Darwinian evolution, Budin
and Szostak (2011) try to identify mechanisms that enabled protocells to
compete with one another. Considering the low levels of membrane
 ARNELLOS; MORENO / FUNCTIONAL DIFFERENTIATION / 7

phospholipids in the early stages of prebiotic evolution, they inquire on

the selective advantage that could have driven the evolution from self-as-
sembled simple single-chain lipids membranes to phospholipid membranes.
They argue that according to their research results “phospholipid-driven
competition could have led early protocells into an evolutionary arms race
leading to steadily increasing diacyl lipid (e.g., phospholipid) content in
their membranes” (ibid, p. 5252).
What is important concerning the quest for minimal FD, is that they
suggest that the resulting decrease of membrane permeability—due to the
increase in membrane phospholipid content during the passage from
self-assembling to self-produced membranes—would impose the cascad-
ing of new selective pressures on early protocells, allowing the internal
evolution of a metabolic and transport machinery that would overcome
the reduced permeability. Specifically, they propose that in order to over-
come the reduced permeability, protocells could have started to evolve
membrane transporters along with proto-metabolic networks for synthe-
sizing their own building blocks, as well as exploring new environmental
niches compatible with compounds that were otherwise rapidly decaying
in fatty acid membranes. They concluded that at this point of transition,
from highly permeable vesicles to less permeable and more stable proto-
cells, a primitive type of selection begins to operate accompanied by a parallel
evolution of the functional domain of the protocells.
Additionally, as also Budin and Szostak mention, the transition from
single-chain lipids to phospholipids had to be gradual. This would allow
for the co-evolution of the metabolic and transport machinery, and would
also permit the overcoming of the initial inefficiency of primitive catalysts.
Accordingly, the primitive operation of the selective dynamics associated
with phospholipid synthesis would correspond to small differences in the
phospholipid content that gradually drive the transition. Considering the
gradual alteration in the degree of the membrane’s permeability, and that
the respective competitive-evolutionary selective dynamics mirror the
emergence of internal metabolic and transport machinery, there should
had been a point where the underlying proto-cellular organization would
had started to produce (though still partly) the main phospholipid com-
ponents of its membrane. What should be noted now is that since the
transition was gradual, the primitive operation of the competitive-evolu-
tionary selective dynamics would had corresponded to small differences
in phospholipid content. Accordingly, the beginning of the internal pro-
duction of the phospholipid parts of the boundary in protocells is a case
where internally generated structural complexity begins to get related to
the dynamics of its production. This point is reached in a developing
protocell only when its membrane permeability starts to decrease and a
proto-metabolic SM network begins to be internally formed and sets about
8 / LUDUS VITALIS / vol. XX / num. 37 / 2012

to operate, thus producing the main phospholipid components of the

membrane. In the evolutionary transition from a self-assembled vesicle to
a self-maintaining protocell, certain structural components and related
processes (the membrane and the self-maintaining chemical proto-meta-
bolic network) begin to play the role of parts in a whole (the self-maintain-
ing organization of the protocell).
Moreover, Budin and Szostak clearly state: “We show that the resulting
increase in membrane phospholipid content would have led to a cascade
of new selective pressures for the evolution of metabolic and transport
machinery to overcome the reduced membrane permeability of diacyl
lipid membranes” (ibid, p. 5249). Accordingly, the resulting increase in
membrane phospholipid content is a kind of ‘control’ of the boundary
conditions, which in turn increases the internal synthesis of phospholipids
residing in the membrane. All this process results in the growth and
consequent division of the protocell (i.e., a primitive mechanism of self-re-
production). Primitive selection dynamics (in this case a phospholipid-
driven competition for single-chain lipids) could thus start operating on
the evolved population of such organizations (a population of protocells)
resulting in a cascading of new selective pressures for several gradual
evolutionary transitions (e.g., emergence of various material structures
constraining different processes, such as efficient transportation of raw
materials, the synthesis of proto-metabolic networks, and so forth). Con-
sequently, the organizational differentiation of the protocells is increased
by recruiting the newly generated (and/or by combining the already
existing) components for new or different organizational tasks (i.e., impor-
tation of material resources, exportation of waste products, exploration of
other niches compatible with structural complexity of a different material
basis, and so on). Overall, this results in an even greater capability for self-
maintenance, hence for persistence.
The evidences from the in vitro experiments of Budin and Szostak,
together with the results of related in silico experiments (see e.g., Mavelli
and Ruiz-Mirazo 2007; Ruiz-Mirazo and Mavelli 2007; Piedrafita, et al.
2009), suggest that the integration of the internally generated (at least
partially) structural complexity in a dynamic organizational frame that
exhibits self-maintenance creates an active separation between system and
environment. This is the point where primitive competitive-evolutionary
selection dynamics can begin to operate.
Initially, the molecular structure of these systems was totally separated
from any form of functionality. Later, the reformed and less permeable
compartment becomes relevant because it makes a contribution to the
maintenance of a proto-cellular organization, in which proto-metabolic
reaction networks are coupled with trans-membrane processes. Indeed, it
is the alteration and the beginning of an internal control in the boundary
 ARNELLOS; MORENO / FUNCTIONAL DIFFERENTIATION / 9

conditions that favors the emergence of organizational complexity, which

is explicitly related to the passage from a self-assembled vesicle to a
self-maintaining protocell. It is in this frame where new molecules, pro-
duced within the protocell, can contribute differently in the maintenance
of the protocell’s organization and then one could speak of FD.
In the next section, we attempt to provide a detailed account of the charac-
teristics, requirements and conditions for the emergence of minimal FD.

3. AN ACCOUNT OF THE ORIGIN OF FUNCTIONAL DIFFERENTIATION

The in vitro experiment analyzed in the previous section shows that certain
SM systems could generate distinct structures contributing in different ways to
their self-maintenance. The system produces different and localizable proc-
esses or structures, each making a specific and distinct contribution to the
conditions of existence of the whole organization. Then, according to
Mossio, et al. (2009), these components become candidates for functional
ascriptions because, in addition to this differentiated contribution to the
SM of the system they belong to, they have been generated (at least
partially) and they are maintained within and by the organization of the
system.
As we have seen, it is likely that in certain environments, chemical
evolution lead to the appearance of populations of protocells with a
significantly lower organizational complexity than the present-day cell,
but which exhibit minimal FD and enable a primitive selection process.
Simple protocellular systems may begin to couple internal chemical reac-
tions with the formation of their membrane components, resulting thus
in the alteration of their boundary conditions, which in turn will allow
them to persist. In that case, the competitive-evolutionary dynamics of a
primitive type of selection can begin to operate giving rise to a mutual
reinforcement between FD and selection for self-maintenance. What is
required for this primitive selection process to be operative is a population
of SM self-reproducing systems that will contribute differently (at least in
a minimal way) to the underlying self-maintenance in which they are
operationally integrated.
Let us discuss the logic leading to the appearance of FD. As mentioned
in section 1.1., in certain SM systems there are components that can operate
as constraints, i.e., their causal role is mainly to modify the underlying
dynamics of the system. For example, certain molecules can catalyze
specific reactions, thus speeding the rates of chemical processes; or they
can self-assemble constituting vesicles, which harness the process they
enclose by ensuring high concentrations. In the scenario described so far,
these structures harness certain processes. Now, in turn, what happens is
that the persistence of these constraints depends on the processes they
10 / LUDUS VITALIS / vol. XX / num. 37 / 2012

harness in an organizationally closed manner. In this way, there is a SM

organization that ensures their repeatable effect, and which results in the
persistence of the system. In a more formal description, we consider a SM
organization as constituted by a set of structures (C1,...,Cn) operating as
constraints. Since in this context each constraint is not able to achieve
self-maintenance on its own, the persistence of a constraint depends on
the process it harnesses, and hence, there should be a form of mutual
dependence between the set of constraints operating in the organization.
In other words, (C1,...,Cn) should constitute “a web of structures exerting
mutual constraining actions on their boundary conditions, such that the
whole web is collectively self-maintaining” (Mossio and Moreno 2010, p.
276). As such, each Ci (as well as any group of constraints) contributes to
the maintenance of the web of the rest of the constraints which, in turn,
maintains in a recursive way (at least some of) the boundary conditions
necessary for the re-appearance and/or maintenance of Ci. We will call this
form of SM organization a closure of constraints (SMCoC) because it is main-
tained due to the action of these constraints and, in turn, these very
constraints are recursively re-produced and/or maintained by the system.
Since the constrained processes generated and maintained in the system
make a distinguishable contribution to its maintenance, they become candi-
dates for differentiated functional ascriptions. It is therefore the operation
of material components as parts of a web of constraints that contribute differently
to the self-maintenance of the system by mutually contributing to the maintenance
(of the boundary conditions) of other constraints in the web that will eventually
classify them as functionally different 1.
In more detail, and according to the arguments for the persistence of
protocells presented in section 2, in order for a system with an organiza-
tion of type SMCoC to be able to interact 2 with the prebiotic environment
and simultaneously to further maintain (i.e., to persist) itself under the
operations of primitive selection dynamics, the following conditions
should be met (see also figure 2):
Cond_1. The operational availability of a subset {Ci} of the set of cons-
traints (C): {C1,...,Cn} operating in SMCoC should go on the constant
formation and maintenance of the boundary conditions (Bd) that will
ensure the continuous operation of the subset {Cj} of the initial set (C),
where i≠j. Then, {Ci} acts so as to determine (at least some of) the Bd_{Cj}.
Cond_2. The operational availability of the second subset of constraints
(Cj) operating in SMCoC under Bd_{Cj} should act so as to determine,
(at least partially) the boundary conditions needed for Cond_1 to be
recursively 3 met (Bd_Cond_1).
Cond_3. {C1,...,Cn} should be, at least partially, internally produced.
 FIGURE 2. Abstract representation of a system exhibiting minimal FD.
A protocell whose organization is maintained by the dynamic coupling of a
chemical SM network, constrained by locally active constraints of the subset {Ci}
(e.g. enzymes, energetic intermediaries, etc.), to the membrane encapsulating
the SM network and acting as another subset of globally active constraints
({Cj}—e.g. playing the role of global boundary conditions that interface the
selective dynamics of the environment and thus alters the physicochemical environ-
ment in (i.e., the Bd_{Ci} under) which the chemical SM network operates and
produces, at least partially, the phospholipid component of the membrane. This
is a special type of SM organization formed on the minimum closure of con-
straints (SMmin[CoC]).

Thus, in a system that operates by simultaneously satisfying all three

conditions there are at least two distinct contributions (realized through
{Ci} and {Cj}), the respective processes of which:
i. participate in the organization of the system that interacts with the
environment, and they contribute to the maintenance of that organi-
zation and consequently, of the respective constraining components
themselves (i.e., there is a SM organization formed on a closure of
constraints - SMCoC);
ii. the system itself generates (partially, at the beginning) those two
subsets of components constraining two processes that contribute
differently to the maintenance of the organization SMCoC (i.e., the
organization of type SMCoC can be minimally differentiated 4).
Therefore, this system exhibits minimal FD and has a SM organization
formed on the minimum closure of constraints (SMmin[CoC]). These conditions
can be realized in a protocell that begins to internally generate the basic
components of its membrane, thus altering the boundary conditions of its
12 / LUDUS VITALIS / vol. XX / num. 37 / 2012

internal chemical network in a way that facilitates its self-maintenance. As

suggested elsewhere (Piedrafita, et al. 2009), this may be the result of the
coupling of two formerly independent systems (self-maintaining chemical
networks such as relatively simple autocatalytic cycles, and self-assem-
bling vesicles that could have been the result of the formation of supra-
molecular structures in a prebiotic environment), that when coupled
together get transformed, becoming strongly inter-dependent operating
under Cond_1, Cond_2 and Cond_3.
The determinative point of the emergence of minimal FD is the begin-
ning of the partial internal generation and coupling of the processes
constrained by {Ci} and {Cj}. As explained in section 2, this should have
been a gradual transition, where the operation of primitive selective
dynamics associated with phospholipid synthesis (e.g., constrained by the
operation of {Ci}) would correspond to small differences in the increase
of the phospholipid content on the membrane of the protocell (e.g., {Cj}).
It is difficult to tell which kind of constraints are required for a minimal
SMCoC, although recent studies in the evolution of protocells show as more
probable the case where sets of structures of different nature have been
integrated by exerting constraining actions on a SM organization. As the
experimental in vitro and in silico results presented and discussed in section
2 show, at least for the moment, the persistence of any SM organization in
a highly changing prebiotic environment seems unrealizable without the
existence of a compartment. Therefore, in the protocells case discussed in
section 2, we could say that {Ci} and {Cj} are two groups of material
constraints of a different type (e.g., membranes playing the role of com-
partments and catalysts).
In particular, the operation of {Cj} interfaces the selection dynamics of
the environment (SEL_env) and thus alters the environment (Env) or, in
other words, partially determines the boundary conditions in which {Ci}
operates. On the other hand, the operation of {Cj} is based on the opera-
tion of {Ci} on the dynamics of a chemical SM network (CHEM_sm) in
SMmin[CoC]. {Ci} partially determines {Cj} and thus affects the way the latter
interfaces the SEL_env, which results in the alteration of the Env in
SMmin[CoC] (see also figure 1).
The subset of constraints {Ci} and {Cj} are operating on two different
aspects with respect to the maintenance of SMmin[CoC]. Specifically:
— {Ci} operates on the dynamics of CHEM_sm and as such it (at least)
partially determines Bd_{Cj}.
— {Cj} interfaces the SEL_env thus making {Ci} operative by (at least)
partially determining Bd_{Ci}.
Therefore, the two subsets of constraints are dynamically coupled but
contribute differently to the maintenance of SMmin[CoC]. We suggest that
 ARNELLOS; MORENO / FUNCTIONAL DIFFERENTIATION / 13

this is a genuine case of a minimal closure of constraints. Accordingly, a

system with an organization of type SMmin[CoC] exhibits minimal FD.
As explained in Budin and Szostak’s experiment, the organization of
type SMmin[CoC] is considered sufficient to initiate a cascading of gradual
evolutionary transitions (emergence of efficient transport mechanisms
and proto-metabolic networks) that would had mirrored the gradual
transition from a partial generation and control of the global boundary
conditions to the almost total internal generation and control of the
diverse ways the system can regulate the exchange of matter and energy
with the environment (e.g., as is the case in modern cells). This will result
in the increase of FD. Let us see how this idea applies to the protocell
example we have analyzed so far (see also figure 3). Let us consider:
a. that with respect to SMmin[CoC], the dynamics of the membrane are, at
least partially, operationally coupled to the chemical dynamics of the
SM network,
b. the gradual alteration in the degree and the type of membrane’s
permeability and
c. that the respective competitive-evolutionary selective dynamics mi-
rror the emergence of internal metabolic and transport machinery.
There could had been a point where certain building blocks inserted in
the membrane (i.e., lipids and peptides) would start to open channels
when the osmotic pressure reaches a certain threshold, thus constituting
precursor channels to control matter inflow/outflow and the release of
waste products. In that case, the organization of the system begins to
evolve from type SMmin[CoC] to type SM[CoC]. Now, the following conditions
are met:
Cond_1. The operational availability of a subset {Ci} of the set of cons-
traints (C1,...,Cn) operating in SMCoC should act so as to determine (at
least some of) the Bd_{Cj} and Bd_{Ck} (i.e. the structures operating
on the formation of primitive channels/transport mechanism) (where
i≠j≠k).
Cond_2. The operational availability of the second subset of constraints
of constraints {Cj} should operate so as to (at least partially) determine
Bd_{Ci} and Bd_{Ck}.
Cond_3. The operational availability of the third subset of constraints
{Ck} should operate so as to (at least partially) determine Bd_{Ci} and
Bd_{Cj}.
Cond_4. {Ci}, {Cj} and {Ck} should be, at least partially, internally
produced.
In other words, the new higher level of organizational complexity of the
system would allow, through its interaction with a dynamic environment,
14 / LUDUS VITALIS / vol. XX / num. 37 / 2012

the generation of distinguishable structural components or processes (e.g.,

primitive channels/transport mechanism) to be differently linked to some
specific results affecting the global operational dynamics and architecture
of that system, with respect to its capacity for even more robust self-main-
tenance in that environment.

FIGURE 3. Abstract depiction of an SM organization exhibiting FD formed on

a closure of constraints SM[CoC].
A protocell whose organization is maintained by the dynamic coupling of a
chemical SM network, constrained by locally active constraints of the subset {Ci}
(e.g., enzymes, energetic intermediaries, etc.), to the membrane encapsulating
the SM network and acting as another subset of globally active constraints ({Cj}
—e.g., playing the role of global boundary conditions that interface the selective
dynamics of the environment and thus alters the physicochemical environment in (i.e.,
the Bd_{Ci} under) which the chemical SM network operates and produces, at
least partially, the phospholipid component of the membrane. Now, the con-
tinuous production and residing of more phospholipid components in the
membrane of the protocell will make it much less permeable, which in turn will
implicitly apply new selective pressures that could result in the formation of
primitive transport mechanisms (e.g. the group of constraints {Ck}) operating
on the release of osmotic pressure, which will at least partially determine the
maintenance of the boundary conditions needed for the continuous operation
of the processes constrained by {Ci} and {Cj} (e.g. avoiding osmotic burst in
the protocell).
 ARNELLOS; MORENO / FUNCTIONAL DIFFERENTIATION / 15

This is a system with a SM organization exhibiting FD formed on a

closure of at least three subsets of constraints SM[CoC]. However, as we shall
discuss in the next section, there is a limit in this process of creation of new
FD. How, then, the organization of a system of type SM[CoC] can evolve
towards significantly higher levels of complexity?

4. TOWARDS HIGHER5 LEVELS OF F.D.

In our analysis, two structures could be functionally different in a system
if the processes that they constrain make different types of contributions
to the maintenance of the system. We have considered that a minimal form
of FD appears in certain SM prebiotic systems, based on a closure of
constraints. These constraints are certain types of molecular structures.
Presumably, this would require, as mentioned, relatively complex molecu-
lar components; yet if compared with present-day biological macromole-
cules, like DNA or proteins, these components should have been much
simpler. Now, for achieving wider repertoires of functional diversity 6,
increasingly higher levels of structural complexity are necessary, since in
chemical SM networks the development of more precise and diverse
functions is directly related to the development of stereo-specificity
(namely, the particular spatial configuration of the components of a big
molecule, which in turn requires the increase of size and composite
variety).
The generation of structural complexity is a spontaneous process of
combination of simpler building blocks as the chemical domain allows the
recombination-based production of structures. Macromolecules and
macromolecular aggregates (supra-molecular structures) are very large
structures (in terms of molecular sizes) held together by intermolecular
forces rather than covalent bonds. Macromolecular aggregates can be
achieved by self-assembly of the molecules that made up the aggregates.
Self-assembly has allowed the construction of a variety of different shapes
and sizes, due to the formation of the aggregates that is directed through
non-covalent interactions. The big difference in the strength of these two
types of bond—covalent bonds hold together material structures by ener-
gies around 1.70 ev, whereas non-covalent bonds are around 0’5 mev—ex-
plains that macromolecular aggregates can be seen as quasi-combinatorial
entities. By this we mean a scenario where basic units can be combined as
building blocks in constructive operations, namely, that parts can be
associated in different ways forming a great variety of stable aggregates,
instead of being driven deterministically towards unique aggregates.
Since the forces that hold together these parts are relatively weak, the
presence of constraints can easily drive the formation of, so to speak,
almost free combinations of building blocks.
16 / LUDUS VITALIS / vol. XX / num. 37 / 2012

Nevertheless, in the conditions of a prebiotic environment the genera-

tion of structural complexity is based mainly in the formation of new
composite molecules, whose building blocks are held together through
covalent bonds 7. This process is limited in size and variety since, among
other reasons, the combinations between building blocks are driven by
specific laws that forbid many possibilities. Now, for the building blocks
to be amenable to combinatorial processes, it is necessary that the dispo-
sitions between the parts be undetermined by intrinsic, lawful interac-
tions. This is the case of structures based on many weak interactions (van
der Waals forces, hydrogen-bond formation, medium-range electro-
static/ionic forces, hydrophobic effect...) acting in parallel within the dif-
ferent building blocks, which allow easily to a considerable number of
combinations among them.
These interesting macromolecular structures can be generated by self-
assembly processes (Lehn 1995), although it is very unlikely that we could
find such type of complex structures in abiotic conditions. Since these
complex structures are formed using relatively weak chemical bonds,
various physical forces can easily damage them. Macromolecular struc-
tures tend to be highly vulnerable to disruption by impinging signals and
are constantly in need to replenishing their dissipating energy and order
as their structural bonds have energies measured in electron volts, even
fractions (Collier and Hooker 1999). Although some relatively complex
structures could be maintained due to spontaneous processes of self-as-
sembly compensating their decay, most of them (as we see in present day
cells) can only be maintained through operations of repair and reproduc-
tion. Hence, though initially relatively complex structures could be formed
in different environments, driven by geological or other types of abiotic
processes, an accumulative production of complex structures leading to
the appearance of interesting macromolecular structures is not likely to
occur outside the frame of SM systems. The reason is that within SM systems
these increasingly complex structures could be recruited to perform func-
tional activities and, therefore, by entering in a SM organization of mutu-
ally dependent constraints, they ensure their maintenance. In other
words, the formation of new and more complex structures can be ensured
if they become functional parts within SM organizations. Thus, although
a large part of biological organization depends on complex structures
formed by self-assembly processes that do not involve energy dissipation
(thermodynamic equilibrium), the variety of functional processes of the
cell results from a combination of complex self-assembly and self-organ-
izing processes (Karsenti 2008).
As we shall see, the consequences of this mutual interplay between
self-organization and self-assembly is the emergence and maintenance of
new larger structures, made of simpler functional components. The cen-
 ARNELLOS; MORENO / FUNCTIONAL DIFFERENTIATION / 17

tral point is that the increase of structural complexity leading to new

functions in a system would be mainly the result of a compositional
processes of the arrangement of the “primitive” functional parts, namely,
the differentiated constraints involved in the organizational closure of the
system.
Through several processes of self-assembly, the resulting structural
complexity can be capable of triggering new structural changes, and new
possibilities for a subsequent composition in an open way of recombina-
tion (see endnote 10). As E. Fox Keller (2009) has pointed out:

The formation of the covalent and non-covalent bonds that hold such molecu-
lar complexes together can also sometimes change the structure of the compo-
nents with which the process started. In so doing, they can also induce changes
in the rules of engagement, thereby creating the possibility for new interac-
tions, new binding sites, new hooks. The new binding sites are not simply the
consequence of the new proximities created by molecular binding, but more
interestingly, of the changes that have been triggered in the ways in which the
component parts can interact. They might be thought of as Brownian motors
in evolutionary space, feeding on chance events to build ever more complex
configurations (...) The phenomenon I am trying to describe rests on two basic
facts: first, that many complex macromolecular structures are capable of stabi-
lizing in a variety of distinctive shapes or forms, and second, that the binding
of new molecules can trigger a shift from one conformation to another, thereby
exposing new binding sites, and new possibilities for subsequent composition
(ibid, pp. 22-23).

This initiates a new and almost indefinitely open domain of structural

variety. Even so, in order to go in this direction, a certain form of modu-
larity is required. A modular organization is one in which different func-
tional components may be separated and recombined, such that allows an
almost free compositional process. Of course, to achieve such result it is
necessary to start with relatively stable entities amenable to a variety of
physical forms of assembly or aggregation. Otherwise, they could not act
as building blocks for the construction of new functional entities. Once
this condition is fulfilled, these basic components can be treated as build-
ing blocks that can be re-combined through processes of arrangement, so
that an open world of combinations can be achieved 8. This way, new more
complex structures can be generated through compositional processes of
arrangement of similar core structures.
This process is interesting since it allows, in turn, the generation of
composite devices able to perform highly specific types of work, namely,
machines. A (mechanical) machine is a structure made of organized parts
that, by its movements, is able to perform a type of work. For example, in
the case of proteins, secondary structures are organized in rigid parts,
18 / LUDUS VITALIS / vol. XX / num. 37 / 2012

whose relative movements can generate different effects, leading to the

amplification of small displacements, or other mechanical effects de-
scribed by words borrowed from the depiction of machines: lever and
spring, ratchet and clamp and so forth (Morange, personal communica-
tion, 2008). To see them as machines performing work requires some input
of energy, which, through a stable arrangement of a set of structures with
specific shapes, is harnessed to perform well specifiable functional processes.
In the case of the ATPase, for example, its chemical function involves a true
mechanical action, as it transports solutes against a gradient. Actually, this
protein works as a small motor, with a rotor and a stator, the first spinning
around the second to generate ATP.
Actually, living cells are constituted by a host of molecular machines
that participate reciprocally in their respective processes of fabrication,
maintenance and operation. And this impressive diversity of functions is
ultimately the support of further levels of functional complexity in living
organisms. This leads us to the main point of this section. When compo-
nents act as building blocks in constructive operations, namely, when
parts can be combined in different ways instead of being deterministically
driven towards unique aggregates, a new open functional domain can be
explored. This is precisely what stands behind the concept of machine.
Actually, the idea of machine is functionally linked to the idea of a
mechanism, which could be defined as a set of “entities and activities
organized such that they are productive of regular changes from start or
set-up to finish or termination conditions”, namely, an arrangement of
parts (a structure) “performing a function in virtue of its components parts,
component operations, and their organization. The orchestrated function-
ing of the mechanism is responsible for one or more phenomena” (Bechtel
and Abrahamsen 2005). Hence, since machines are the result of composi-
tional processes of arrangement of parts, these arrangements could be used
to generate new ones, and this process can proceed indefinitely. A given
“simple” machine can in turn be integrated in a more complex arrange-
ment of parts, such that new functions could be achieved 9.
This implies that through processes of arrangement of parts an open 10 world
of functions can be achieved. Since machines are the result of compositional
processes of arrangement of parts, and that the output of these new, more
encompassing arrangements could be also a new function, new functions
could be created. In turn, these processes could be used to produce new
and more complex molecular machines, in a recursive way, and these new
machines can generate new functional tasks. Therefore, machines are on
this perspective, the material support of mechanistic processes which in
turn allow the creation of unlimited functional diversity.
 ARNELLOS; MORENO / FUNCTIONAL DIFFERENTIATION / 19

5. CONCLUSIONS
As argued elsewhere (Moreno and Ruiz-Mirazo 2009), it now seems even
more possible that the source for FD lies ultimately in the organization of
(proto-) biological systems. For two reasons. First, because certain SM
systems formed on a closure of constrains (SMCoC) can reproduce them-
selves and, therefore, support some form of continuity allowing, at least
in principle, an accumulation of novelties. Second, because SMCoC systems
are potentially capable of producing and preserving new components of
increasing structural complexity. In other words, SMCoC systems provide a
framework to allow the production of a wide variety of new components,
some of which could be recruited by the systems themselves for modifying
their own organization. This is precisely the scenario we have suggested
for the origin of FD. Interestingly, the exploration of new forms of struc-
tural complexity (new components) through assembling processes implies
a certain degree of decoupling from the dynamics of the chemical SMCoC
organization. This allows a relatively free exploration of the domain of
structural complexity, which ultimately can be recruited for increasing in
an open way, in turn, the functional complexity of the overall SM organi-
zation. This scenario opens up an unlimitedly wide domain of harnessed
processes. Matter and energy flows can be manipulated to achieve highly
complex and specific tasks. Therefore, by recruiting these macromolecular
machines SM systems can create new and more complex ways of viability
and develop internal subsystems, for example, adaptive mechanisms.

ACKNOWLEDGMENTS
The authors wish to thank Matteo Mossio and Kepa Ruiz-Mirazo for valuable
feedback on earlier versions of the manuscript. Argyris Arnellos holds a Marie
Curie Research Fellowship (IEF-273635). Alvaro Moreno acknowledges the aid
of the Research Project IT 505-10 of the Gobierno Vasco and FFI2008-06348-
C02-01/FISO-FFI2008-06348-C02-02/FISO and FFU2009-12895-CO2-02 of the
Spanish Ministerio de Ciencia e Innovación.
20 / LUDUS VITALIS / vol. XX / num. 37 / 2012

NOTES

1 This has some important implications with respect to many physical or

mechanical systems that are mistakenly considered as fulfilling the condition
of a closure of constraints (CoC), which in turn, contribute differently in the
maintenance of the underlying organization. An example would be a billiard
table where balls are arranged in a way that can hit the one another after
striking on the sides of the table. Assuming a frictionless table, the whole
setup can be maintained for a long period of time thus achieving CoC, and
assuming different forces responsible for the initiation of the movement of
each ball we can have different speeds thus achieving different contributions.
However, it is obvious that we would be very reluctant to discuss about FD
in such a system since the whole closure is achieved through the externally
imposed boundary conditions (speed and space of movement), let alone the
creation of the structural components of the system. Another, more interest-
ing example is Ruse’s (2003) rain-cloud-river water cycle, where a set of
structures (clouds, rain, spring, river, sea, clouds) seems to have mutual causal
relations that fit the idea of FD. For instance, the clouds generate the rain,
which (contribute to) generates a spring, which generates a river, which
(contribute to) generates a lake, which generates different clouds, and so on.
However, this is not the case. Not only the water cycle does not exhibit any
CoC (see Mossio and Moreno, 2010 for an analytic explanation), but also, the
generation of any structure (e.g. a cloud) that may contribute differently (e.g,.
heavy or light rain) to the cycle is the result of externally imposed boundary
conditions (e.g., sun, air, various types of ground), which since there is no
CoC they are not redefined by the operation of the constraints in the system.
Therefore, the water cycle does not exhibit any form of functionality.
2 It is noted that even at this level of organizational complexity the system
interacts with the environment, although its interaction is limited in the
exchange of raw chemical materials.
3 By ‘recursively’ we mean that the fulfilment of Cond_1 will ensure the
maintenance of Cond_2, so that under certain environmental conditions their
operational combination (Cond_1 & Cond_2) will ensure Cond_3, whose
fulfilment will result in the continuous persistence of the operational combi-
nation of Cond_1 & Cond_2 & Cond_3) and so on. At this level of organiza-
tional complexity of a biological system exhibiting minimal FD, the mutual
dependence of those conditions and of the respective constrained processes
is of a recursive nature.
4 At this level of organizational complexity there is no need for the system to be
able to alternate between two significantly different modes of operation thus
presenting different self-maintaining characteristics and properties, but there
is a need for the association of the effects of a set of constrained processes of
the system with at least two significantly different contributions to the overall
SM organization.
5 In this section, we will discuss the prerequisites for significantly higher levels
of FD but we will not necessarily consider open-endedness in an unlimited
way.
6 At this level of organizational complexity, it is postulated that a minimum
degree of functional differentiation (FD) is already exhibited by the system
and hence, what we consider is a possible way for the system to achieve a
significant openness of that differentiation.
 ARNELLOS; MORENO / FUNCTIONAL DIFFERENTIATION / 21

7 There is also the formation of supramolecular structures, like processes of

crystallization, but these bigger structures have poor internal diversity unless
they are based on weaker bonds, as it is the case for lipid membranes, micelles,
polymer aggregation, etc.
8 It is not easy to tell what degree of molecular complexity is required to satisfy
these conditions. But it seems that relatively small ones could act as modules.
For example, as recently shown by Manrubia and Briones (2007), certain small
molecules of RNA can play the role of modules in a stepwise model of
ligation-based modular evolution: RNA hairpin modules could have dis-
played ligase activity, catalyzing the assembly of larger, eventually functional
RNA molecules. These ligation processes allow a fraction of the population
to retain their previous modular structure, and thus, structural and functional
complexity can progressively increase, even in the absence of template repli-
cation.
9 It is very important to keep in mind that the output of these arrangements is
a function, and that a structure of parts is only a machine if it serves a function
in a given system.
10 “Open” here refers to the idea that functional devices could be used to
generate new ones, and that this process can precede indefinitely. Thus,
functional ‘recursion’ (i.e. functional recombination in the context of a SM
organization) is a more powerful concept than that of the aforementioned
assembling, based solely on structural re-combination processes.
22 / LUDUS VITALIS / vol. XX / num. 37 / 2012

REFERENCES

Bechtel, W. and Abrahamsen, A. (2005), “Explanation: a mechanistic alternative”,

Studies in History and Philosophy of the Biological and Biomedical Sciences 36:
421-441.
Bickhard, M. H. (2000), “Autonomy, function, and representation”, Commun.
Cogn. Artif. Intell. 17 (3–4): 111–131.
Bickhard, M. H. (2004), “Process and emergence: normative function and repre-
sentation”, Axiomathes 14: 121-155.
Budin, I., and Szostak, J. W. (2011), “Physical effects underlying the transition
from primitive to modern cell membranes”, PNAS 108 (13): 5249-5254.
Chen I. A., Roberts R. W., and Szostak J. W. (2004), “The emergence of competi-
tion between model protocells”, Science 305: 1474–1476.
Christensen W. D. and Bickhard, M. H. (2002), “The process dynamics of nor-
mative function”, Monist 85 (1): 3–28.
Collier, J. (1999), “Autonomy in anticipatory systems: significance for function-
ality, intentionality and meaning,” in Dubois, D. M. (ed.), Computing Antici-
patory Systems, CASYS’98 - Second International Conference, American
Institute of Physics, New York: Woodbury, pp: 75-81.
Collier, J. (2000), “Autonomy and process closure as the basis for functionality”,
Annual NY Academy of Science 901: 289-291.
Collier, J. (2004), “Interactively open autonomy unifies two approaches to func-
tion,” in Dubois, D. M. (ed.), Computing Anticipatory Systems, CASY’03 - Sixth
International Conference, American Institute of Physics, Melville, New York,
AIP Conference Proceedings 718, pp: 228-235.
Collier, J. D. & Hooker, C. D. (1999), “Complexly organized dynamical systems.”
Open Systems and Information Dynamics 6: 241-302.
Delancey, C. (2006), “Ontology and teleofunctions: A defense and revision of the
systematic account of teleological explanation”, Synthese 150: 69-98.
Edin, B. (2008), “Assigning biological functions: making sense of causal chains”,
Synthese 161: 203-218.
Fox Keller, E. (2007), “The disappearance of function from ‘self-organizing
systems’,” in Boogerd F, Bruggeman J.H., Hofmeyr H.V., Westerhoff H. (eds.),
Systems Biology. Philosophical Foundations. Elsevier, Dordrecht, pp: 303–318.
Fox Keller, E. (2009), “Self-organization, self-assembly, and the inherent activity
of matter,” in S. H. Otto (ed.), The Hans Rausing Lecture 2009 (Uppsala Univer-
sity, Disciplinary Domain of Humanities and Social Sciences, Faculty of Arts,
Department of History of Science and Ideas).
Fox Keller, E. (2010), “It is possible to reduce biological explanations to explana-
tions in chemistry and/or physics,” in Francisco J. Ayala and Robert Arp (eds.),
Contemporary Debates in Philosophy of Biology. NY: Wiley, pp: 19-31.
Glansdorff, P. and Prigogine, I. (1971), Thermodynamics of Structure, Stability and
Fluctuations. London: Wiley.
Godfrey-Smith, P. (1994), “A modern history theory of functions”, Noûs 28:
344–62. Reprinted in D. J. Buller (ed.), 1999, Function, Selection, and Design,
Albany, New York: SUNY Press, pp. 199–220.
Godfrey-Smith, Peter (2009), Darwinian Populations and Natural Selection. Oxford:
Oxford University Press.
Karsenti, E. (2008), “Self-organization in cell biology: a brief history”, Nat Rev Mol
Cell Biol. 9 (3): 255-262.
 ARNELLOS; MORENO / FUNCTIONAL DIFFERENTIATION / 23

Lehn, J. M. (1995), Supramolecular Chemistry: Concepts and Perspectives. New York:

Wiley-VCH.
Manrubia, S. and Briones, C. (2007), “Modular evolution and increase of func-
tional complexity in RNA molecules”, RNA 13(1): 97-107.
Mavelli, F. and Ruiz-Mirazo, K. (2007), “Stochastic simulations of minimal self-
reproducing cellular systems”, Phil. Trans. R. Soc. 362: 1789–1802.
Millikan, R. G. (1989), “In defense of proper functions”, Philosophy of Science 56:
288–302.
Morange, M. (2008), “The increasing place of macromolecular machines in the
in the descriptions of molecular biologists”, personal communication.
Moreno, A. and Ruiz-Mirazo, K. (2009), “The problem of the emergence of
functional diversity in prebiotic evolution”, Biology and Philosophy 24 (5):
585-605.
Mossio, M., Saborido, C. and Moreno, A. (2009), “An organizational account for
biological functions”, British Journal for the Philosophy of Science 60 (4): 813-841.
Mossio, M. and Moreno, A. (2010), “Organisational closure in biological organ-
isms”, History and Philosophy of the Life Sciences 32 (2-3): 269-288.
Neander, K. (1991), “Function as selected effects: The conceptual analyst’s
defense”, Philosophy of Science 58: 168–184.
Nicolis, G. and Prigogine, I. (1977), Self-Organization in Non-Equilibrium Systems:
From Dissipative Structures to Order Through Fluctuation. New York: Wiley.
Nicolis, G. (1989), “Physics of far-from-equilibrium systems and self-organisa-
tion”, in P. Davies (ed.), The New Physics. Cambridge: Cambridge University
Press, pp: 316–47.
Piedrafita, G., Mavelli, F., Moran, F. and Ruiz-Mirazo, K. (2011), “On the transi-
tion from prebiotic to protobiological membranes: from ‘self-assembly’ to
‘self-production’”, in: LNCS, 5777/2011: 256-264.
Ruiz Mirazo, K. and Mavelli, F. (2007), “Simulation model for functionalized
vesicles: lipid-peptide integration in minimal protocells”, in Almeida e Costa
F, Rocha LM, Harvey I, Coutinho A (eds.) ECAL 2007, Lisbon, Portugal,
September 10–14, 2007, Proceedings. Lecture Notes in Computer Science
4648. Springer, Heidelberg, pp: 32–4.
Ruiz-Mirazo, K. and Mavelli, F. (2008), “On the way towards basic autonomous
agents: stochastic simulations of lipid-peptide cells”, Biosystems 91: 374-87.
Ruse, M. (2003), Darwin and Design: Does Evolution have a Purpose? Cambridge,
MA: Harvard University Press.
Segre, D. and Lancet, D. (2000), “Composing life” EMBO Rep 1 (3): 217–222.
Schlosser, G. (1998), “Self-re-production and functionality: a systems-theoretical
approach to teleological explanation”, Synthese 116: 303–354.
Walde, P., Goto, A., Monnard, P. A., Wessicken, M. and Luisi, P. L. (1994),
“Oparin’s reaction revisited: enzymatic synthesis of poly (adenylic acid) in
micelles and self-reproducing vesicles”, J Am Chem Soc 116: 7541–7547.