ecological complexity 4 (2007) 93–101
available at www.sciencedirect.com
journal homepage: http://www.elsevier.com/locate/ecocom
View point
Towards an ecosystem semiotics Some basic aspects for a new research programme Soeren Nors Nielsen * University of Copenhagen, Faculty of Pharmaceutical Sciences, Institute for Pharmaceutics and Analytical Chemistry Toxicology and Environmental Chemistry, Universitetsparken 2, DK-2100 Copenhagen, Denmark
article info
abstract
Article history:
The paper argues that ecosystem should be recognized as semiotic systems and that it is
Received 30 March 2007
necessary to carry out studies of the ongoing semiotic processes in addition to traditional
Received in revised form
ecosystem research. It is suggested that interpretation of ecosystems within such a semiotic
10 April 2007
framework is of utmost importance and essential if we want to fully understand the
Accepted 13 April 2007
complexity issue and how complex behaviour comes about at this level of biological
Published on line 5 June 2007
hierarchy. This area—called ecosystem semiotics—is suggested to become a new direction of study dedicated to this understanding.
Keywords:
As a consequence of the ontic character of ecosystem complexity, studies on the impor-
Semiotics
tance of semiotic processes can only be synthesized through modelling efforts. Hitherto, this
Complexity
type of process with a few exceptions has been neglected or at best only implicitly integrated
Ascendancy
and accounted for in ecosystem models. In the future, ecosystem models will need to integrate
Indirect effects
this type of behaviour in order to get full insight into the causal mechanisms behind the
Emergent property
emergence of their complex behaviour. In addition, the concept of exergy in its classical form
Information
derived by Evans is suggested as a platform to integrate thermodynamic information of the
Ecosystem
systems as a complexity measure. The thermodynamic information may be split into parts
Niche construction
that causally originate in the ontic existence of various ecosystem elements. Ecosystem semiotics is thought to considerably increase the thermodynamic efficiency of the ecosystem, leading to an increase in thermodynamic information and for instance ascendancy that would not have existed if it was not emerging from the semiotic processes. In other words, by incorporating semiotics, we add a ‘‘metaphysical’’ layer to our models, which may be referred to as the semiotype of the system. The semiotype acts as downward causation on the lower layers of interactions and allows for modification and adaptations of existing genotype or phenotype possibilities that would not be possible without the existence of semiosis and cognitive processes. # 2007 Elsevier B.V. All rights reserved.
1.
Introduction
Many colleagues will, upon reading the above title, immediately come up with questions like the following: Does this
mean that ecosystems possess a language and that they ‘‘talk’’? Do they posses some special type of sign-interpreting ‘‘psychology’’? For a start, the answer must be as follows: Well, at least in the eyes of this author and a few others—to a certain
* Tel.: +45 35 30 64 55/60 00/62 61; fax: +45 35 30 60 13/55/01/10. E-mail address:
[email protected]. 1476-945X/$ – see front matter # 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.ecocom.2007.04.001
94
ecological complexity 4 (2007) 93–101
extent, and in a certain way, they do. But let us see if we can get closer to a more accurate answer or a more correctly formulated question. To begin with, many of our colleagues would probably find it easier and more acceptable to look at ecosystems as some sort of communication system, so let us take this as an entrance point to this treatment. In fact, in ecology, the use of equations derived from communication theory has found a widespread use in studying the phenomenological behaviour of ecosystems. Just think of the implementation of the Shannon–Weaver or similar information indices in studies of bio-diversity (Marques, 2001). The application of the measure of average mutual information as an index of the developmental state of an ecosystem network is another, to mention but a few, clear examples (e.g. Ulanowicz, 1986, 1997). So, in order to proceed and take these approaches a step further into the study of ecosystems, let us end this discussion for now by stating: at least nature and ecosystems communicate somehow. Having said this, in the rest of this paper, we put the focus on defining and understanding this communication and its importance to the emergence of complex behaviour. Although, it may sound strange, it may well be that the widespread use of information indices has prevented us from taking exactly this step, in a semiotic direction. We have been too much used to automatically apply the above-mentioned types of diversity equations in the former ontological contexts to see the possibilities of applications in other areas. This is somewhat peculiar, as it is to a certain extent only the full consequence of what we are already doing. Life is full of and almost defined by semiotic processes (Hoffmeyer, 1996, 1998; Emmeche, 1998). As ecosystems are composed of life forms, why not extend the semiotic view to ecosystems. Put in other, simpler words, why do we apply information indices to ecosystems if we do not at the same time accept them as information processing, i.e. semiotic, systems? ‘‘Information processing’’ leads to constraints, and as Rutledge et al. (1976) showed, are ripe candidates for quantification using information theory. A first attempt of combining semiotics to informational measures and indicators of processes in ecosystems may be found in Ulanowicz (2002). By the way, what exactly is semiotics? For simplicity, we will here adopt the definition by Umberto Eco that ‘‘semiotics is concerned with everything that can be taken as a sign’’ (quoted from Chandler, 2002), meaning that semiotics is the study of signs and sign-interpreting systems. This immediately leads to the question: what is a sign? Again to take one of the widest definitions, according to Peirce ‘‘A sign is something that means something to somebody in some context’’ (my translation). By adopting these wide definitions things may not immediately become more, but at the end it will facilitate the transfer of the view to other areas. To extend this view to ecosystems will not be an easy task, because based on recent studies we are just starting to get a picture of what is going on among all the ecosystem components, the processes that connect them and how they are controlled. But exactly this increase in insight is maybe making us ready to discuss how further progress can be made. A logical name to this approach of extending the semiotic view to ecosystems would be semiotic ecology, ecological semiotics
or eco-semiotics for short. Meanwhile, these terms have been used already in another context basically covering the relationship between culture and nature (Kull, 1998). Biosemiotics is according to Emmeche et al. (2002) ‘‘biology that interprets living systems as sign systems’’—insofar that we consider ecosystem to be collections of biological, living subsystems, we, by addressing the ecosystem semiotics, take biosemiotics to a metaphysical level. We may call it the semiotics of the super-system, but not in the sense of a superorganism, thereby indicating that we deal with biosemiotics with only a very light Clementsian ‘‘flavour’’ added to it. Super is here referring to pre-syllable meta in the sense of ‘‘higher level’’ rather than in the sense of superior which would indicate a higher quality. The latter remark is probably not a philosophically correct statement but may serve to facilitate to convey the message of what ecosystem semiotics will be. For further treatment of biosemiosis, please refer to Emmeche and Hoffmeyer (1991). Another reason for the preference of the term ‘‘ecosystem semiotics’’ is that the eco-semiotics of Kull (1998) brings human culture to the outside of the ecosystem. True ecosystem semiotics should include humans as components, representing the first and the second nature as defined by Von Uexku¨ll (1926). The point to be made in this paper is that the communicative, i.e. semiotic, processes in ecosystems and the role they play in maintaining important functions tend to be an overlooked area. Thus, the information issue has hitherto been addressed based on ontologically existing ecosystem elements like species or organisms or the material and energetic fluxes between them. This may have been a shortcut in the analysis, but it tends to ignore the role of communication and sign interpretation in shaping and controlling the fluxes as well as the elements. This is an imperfect approach as ecosystem semiotics seems to be an area of utmost importance for understanding the complexity issue of ecosystems, both the complexity in composition and niches, as well as in behaviour, adaptation and evolution. For a thorough treatment of the complexity issue, see Emmeche (1997).
2.
Information and biological systems
As described in an earlier contribution, which attempts to analyse the use and misuse of information-theoretic approaches in the analysis of ecosystems (Nielsen, 2000), the inevitable ‘‘takeover’’ of problems already inherent in the concept of information was discussed. One major problem, valid also in this context, includes an almost classical debate on whether information may be viewed as existing even under passive storage, or if information can only be dealt with and understood properly when transmitted or conveyed. Furthermore, does this conveyance need to be successful or not? In other words, does a library contain information just because it consists of books? Or do they need to be read or even understood for the information contained in the books to exist? Next, after one has decided on the possible ontological existence of information together with its actual character, one needs to come to an agreement on how to measure it. A
95
ecological complexity 4 (2007) 93–101
variety of measures claimed to express the information of a system sensu lato have been presented in the scientific literature since the middle of the last century, part of which, as seen from the introductory section, have also been introduced in ecological science (Jørgensen and Mejer, 1981; Ulanowicz, 1986, 1997; Shannon, 1948, 1951; Bennett, 1988). It might in the end turn out to be not so important which measure to use when analysing the temporal and spatial development of one particular system. But when it comes to a situation of comparison and evaluation of several systems against each other, for instance in terms of complexity, it becomes of utmost importance that the expressions used are consistent, for instance in what base of logarithm is used. This
consistency is also necessary to carry out the exercise in and objective manner. This means that using expressions that are mathematically isomorphic may not be a sufficient criterion for the validity of a comparison. In addition, the derivation, i.e. ontological background of any particular expression, needs to be taken into consideration, as demonstrated in the following. Last but not least, the relation between physical and mathematical concepts is unclear and therefore information has often been linked to the concepts in a vague and inaccurate manner. From time to time, concepts have been taken out of their original domain and used in a reverse manner compared to how the concept was intended. One famous example is the concept of (information) entropy
Table 1 – A traditional scalar hierarchy of the system Level of hierarchy Molecular Pre-biotic and biogeochemical systems Biochemical level Hormonal level
Cell level Cell nucleus
Cell organelles
Organ level Neuronal system
Brains
Organism level Individual Populations, societies
Ecosystem level Far from equilibrium structure
Network structure
Complex systems theory
Examples of important process Autocatalysis minimum free energy
Autocatalysis, hypercycling and centripetality Female hormonal system (pheromones)
DNA replication and transcription into m-RNA
Examples of semiotics Chemical bondings, primary, secondary tertiary and quaternary structures
Information measures Basic thermodynamics
Do Transmitters and receptors— ‘‘key in loc depict ions
Proofreading reading of triplet codes (vertical communication or genealogical semiosis, Hoffmeyer, 1996)
‘‘Atomist’’ view of information binary at the extreme
Protein production translation into molecular signifer
Do
Exchange of transmitters, causing electric signals to progress calcium and other ions Integrated, coupled linking of sign in complex network of neurons
Analogue systems
Coupling of all major (physiological processes) Intra-species communication, extra-species communication
Self-organised, dissipative superstructures exchange via thermodynamic relations Focused on fluxes and feedbacks (controls), exchange of energy and matter as information exchange
The whole new area
Analogue to binary
Synergistic and interpretory behaviour Pheromones, ethological process
Both analogue and binary
Hierarchical and far from equilibrium thermodynamics 2. Law oriented Average mutual information Network thermodynamics 1. Law oriented Logical depth and algorithmic complexity—binary
At each level, examples of important processes are indicated together with the semiotic process most commonly involved. At the end, some commonly implemented views or approaches to interpretations within an informational framework are also exemplified. Information is here treated in the widest sense, spanning from pure information theoretical approaches to implementations of concepts such as thermodynamic information concept. The list is not intended to be exhaustive only giving examples of serving the purpose of illustrating the ideas sketched in the paper.
96
ecological complexity 4 (2007) 93–101
presented by Shannon and Weaver (1949) as opposed to the concept of entropy in thermodynamics as derived from Boltzmann’s and Gibb’s equation(s). The equations of both entropies are isomorphic but tell quite different things about the ontics of the system. Exactly this ontic relationship is probably the most likely cause of confusion among biologists when discussing information. It is caused by the fact that researchers in a particular area will tend to pick out the view of information which fits the best into their ontic views and will be most likely to receive less resistance to other researchers within their paradigmatic area. The major importance of this situation may be seen from Table 1. Glancing through literature, no conscious patterns seem to exist in the choice of authors to any of the equations, rather the choice seems to be determined by tradition and history, i.e. paradigm. On the importance of this to semiotic interpretations, see later.
3.
Information and noise—a classical view
Information content, in particular when transmitted or conveyed, is subjected to the introduction of noise. This was originally the problem under analysis by Claude Shannon, investigated for the English language. An example of a typical information-theoretic view that may serve to illustrate the issue of noise in communication processes is found in Fig. 1. Here, a radio signal is transmitted from an antenna to a radio containing a receiver and a loudspeaker, through which the original message sent out by the radio station is further conveyed to a potential listener. Noise may be introduced at numerous places in this system: malfunctions in the transmitting antenna (a), atmospheric disturbances (b), malfunctions in the receiving antenna (c), the parts transforming and transferring the signals (d) and finally in the loudspeaker itself (e), interference from other sounds in the surroundings of a listener, which may receive the signal in a disturbed manner due to physiological (g) or psychological (h) reasons, etc. The
Fig. 1 – Phases of communication. The figure illustrates the possibilities of malfunction in the last stages of a radio broadcast, that is the from transmission of radiowaves to sound eventually enters the ear of a potential listener. Letters indicate possibilities of malfunction and noise to enter the process, please refer to text.
possibilities are so many that it ought to make us wonder how intentional information transfer is possible at all. While this classical view focuses on the transfer of information, one must not forget that information storage plays an important role and is most often a prerequisite to a successful transmission, for instance storage may be necessary between recording and broadcasting, also even if we think of it as a direct transmission. Broadcastings are most often stored on tapes, DVDs or other media with the purpose of reuse or in order to save recordings of events for history. When adding this perspective to communication theory, we can eventually get very close to a biological communication system. In nature, information is, e.g., stored in our genotype and used, being expressed, as we prefer to call it, during various phases during our life cycle. The expression is thought to be the basic platform for the function of our system (see later on geno-, pheno- and enviro-type view).
4. Information in environs—or ecosystems interpreted as semiotic systems The transfer of this view of transmissions and disturbances to biological systems may be easily understood when coupling it to the concept of the environ, introduced by B.C. Patten (1978) and corresponding to the works of the Estonian animal physiologist and ethologist Jacob Von Uexku¨ll (1926). In a simplified explanation according to the views of Uexkiill, an organism will receive stimuli from the surroundings, Umwelt, by Patten called the world of perception or world as sensed. The organism would process the signals in its inner world, and make a proper, adequate response to the received stimuli. This would take place through an interaction with the surroundings, in this case called the world of action. The whole would have been a matter of a simple action and reaction scheme had it not been for the fact that von Uexku¨ll added a possibility of an interaction between the two outer worlds. Thus the world of action may feed back and have an impact on the world of perception. This happens through what is called a function cycle. An example would be a male blackbird in spring that receives stimuli (from the world of perception) carrying information that the period of day light increases, temperature rises and trees start to sprout. As a response (to its world of action) it will start to sing, find a suitable habitat or territory, and attract a female, eventually representing a new stimulus. Meanwhile, quality matters. The more ‘‘beautiful’’ or intense the male is singing, the more safe the chosen habitat is, the more aggressive the behaviour and success in keeping intruders away, the more likely it is to get the ultimate reward, to establish a couple and produce offspring. Patten (starting in 1978) transfers the view of von Uexku¨ll to our understanding of ecosystem structure and function. An ecosystem may be seen as, what is now referred to, an environ (a within-system analogue of the outer world of von Uexku¨ll, and extension of Koestlers principle of the Holon as an organisational unit). It corresponds to the world of perception, an ‘‘upstream’’ component called the input environ, through which it may receive signals, usually in the form of material or energetic flows from the outside. A component of the ecosystem or the components together process the signals
ecological complexity 4 (2007) 93–101
and pass them on to other elements or the ‘‘downstream’’ component, the output environ. Feedbacks are taking place from the output environ to the input environ through the function circle. As a consequence of these interconnections, the ecosystem becomes complex in structure and behaviour as expressed in ‘‘20 remarkable properties’’ by Patten (1998). Remarkable seems here also synonymous with emergence, cf. Nielsen and Mu¨ller (2000). It is now easy to draw a parallel between the noise in our transmission system and the disturbance of the communication processes of the organismic system, the blackbird or an ecosystem (see Fig. 2). It is easily seen that malfunction of ecosystems may be introduced by total or even only partial interruption or malfunction occurring in input or output environs, the environ itself, or in the function circle, i.e. in four principally different ways. Now, as we shall see when dealing with ecosystems, the situation gets more intricate and complex. First, of all this is caused by the fact that when analysing noise in a transmission system (as in Fig. 1), we tend to deal with one particular type of noise, in the above example case interference on radiowaves is more or less sufficient to understand how malfunctioning comes about. Meanwhile, ecosystems, as shall be seen in the following, have more ways of transferring ‘‘signals’’ (sensu lato) and their behaviour may be dependent on all of them, or the relations between them in a rather unclear manner.
Fig. 2 – Systems as environs. Part (a) represents the normal way of depicting the environ concept of B.C. Patten (1978), notation being Z for input environ, X for the environ itself and Y for the output environ. Part (b) represents an analogue to Fig. 1, where fat dotted lines indicate possibilities of errors in transfer between compartments.
97
The input environ supplies stimuli of both quantitative, material (or often materially bound energy) character and qualitative, non-material character. The environ will treat both types of stimuli and forward responses of both the quantitative and qualitative character through the output environ. In this context, the input environ becomes a causal, logical sequence of transfers, treatments, and transformations of information through which the output environ may propagate impulses back to the input environ through the information network. Viewing all the transfers as ‘‘informational’’ allows us to see that quantity and quality are, in general, not separable in the ecosystem but highly interdependent and of a very complex character. This is not exclusive! Meanwhile, some separation is possible and does make sense such as the separation of a quantity like system size expressed as total system throughput (TST) from the distribution of this quantity, which for example finds its measure in the average mutual information of the system (AMI) as suggested by Ulanowicz (1986, 1997). There is not necessarily a one-to-one relationship between stimuli and response in between quantity and quality (with almost a guarantee there is not!)—quantity may be changed into quality and vice versa. Noise may now be introduced, but mechanisms through which interactions take place are so intricate that it may easily be seen that the resulting system behaviour must be complex. The point is now not only that information and communication within the ecosystem play and important role in raising its complexity and creating its complex behaviour. Hitherto, ecological researchers have tended to focus on the quantitative aspect of ecosystems and even today the vast majority of ecosystem models constructed are concentrating on material and/or energetic fluxes. As a result, the possible importance of the qualities of the systems, e.g. the importance of communication or semiotic processes tends to be neglected. It is true though, that studies focusing on the causal role of mechanisms like autocatalysis (Ulanowicz, 1997, 2001) to a certain extent do include an acceptance of the existing underlying semiotic processes of being important to the system. Meanwhile, the studies on the phenomena such as indirect effects, network amplification and synergism carried out by Patten and coworkers (for an overview see Patten, 1998) may be the closest we get today in revealing the possible interaction between material fluxes and the role of semiotics in shaping the system. For instance, the network elements do seem to (not consciously) cooperate in a manner that leads to a conversion of direct interactions into more positive, indirect interaction forms (see Patten, 1998). These qualitative changes in interaction, according to this author, to a wide extent originate in the semiotic layers of the system. The point here is to stress that our studies hitherto may have been far too deterministic to capture the role of semiotics in the controls and cybernetics of the system.
5. The information perspective and biological systems It has become clear from the above analysis that the suitability of the various informational concepts were differing a lot and
98
ecological complexity 4 (2007) 93–101
that this in fact has to do with the level of hierarchy and thus the complexity level of the system. During recent years, other types of information have been developed that are more macroscopic and holistic in character and therefore may be applicable to systems at higher levels, i.e. in domains that are far away from even the ‘‘far from equilibrium’’ domain of thermodynamics. What has been found most suitable here is to take a starting point in the analysis at any level in the formulation of exergy (Ex) after Evans et al. (1966) stating that Ex ¼ T I and I ¼ Seq Sstate where T is the absolute temperature (in K) of the system surroundings, I is the thermodynamic information of the system, Seq refers to the entropy at equilibrium state (by definition maximum) and Sstate refers to the actual entropy state of the system. An additional strength of using this expression is that the ‘‘noise’’ introduced from variation in temperature on the information content will be in the order of 3–4% which is much below the general uncertainties found in connection with ecological experiments. An approach originating in the equation of Evans is appealing, as it is now possible to divide the contribution of various types of information to the deviation of a certain system from thermodynamic equilibrium or any other reference state that might be chosen. In fact, it may turn out to be applicable to any system outside the far from equilibrium domain, i.e. on any level of biological hierarchy.
6. Connecting thermodynamic information and environs In lectures on modelling given by B.C. Patten (personal communication 1995), it is often suggested that the function of any element of a system may be classified as belonging to one of the following identifiers—either compartments, connections or controls—the concept is sometimes referred to as the concept of the three Cs. Compartments are more or less equivalent to what is also referred to as state variables, connections are the flows or processes of the system. Roughly speaking, a control is any other element not at first identified as compartment or connection. Meanwhile, both compartments and connections may at a later stage in the systems analysis and model development be identified as playing the role as a control and may therefore enter the model system as such and exert for instance a positive or negative feedback on any other element in the system. Building on the above concept, it is possible to identify three major types of information that play a part in determining the complexity level of the system, one belonging to the components, one stemming from the flows and one from the system controls. For easiness, they will be treated as additive entities, interpreting the next information in the hierarchy of three Cs as the information added (an extra layer
in the total) to the system by introducing the extra types of elements. Thus, it will be possible to state that: Itotal ¼ Icomp þ Iconn þ Ictrl This may be related to other measures of ecosystem state derived from thermodynamics, information or network theory. Icomp is close to what was expressed by the classical exergy concept introduced by Førgenren and Mejer which was in fact based on the equation defining Evans and Kullback’s measure of information. Iconn is more or less equivalent to simple network measures such as connectivity or connectance and included in the ascendency of Ulanowicz, i.e. probably more related to the average mutual information (AMI) of the network. The Ictrl part now remains to be explained or interpreted, although somehow already embedded in the network organisation and structure. For other ways of splitting the information, in particular flows, in ecosystems please refer to Ulanowicz and Norden (1990), Ulanowicz and Abarca-Arenas (1997) and Nielsen and Ulanowicz (2000). Referring to these considerations studies characterizing the importance of Ictrl must be at the core of ecosystem semiotics as control exerted by semiotic processes and nonmechanical in character will be a subset of Ictrl. Ictrl is then an informational fraction which has not really yet been covered by any of the relevant ecosystem theories or at least not by the vast majority of the models we make. Therefore, new attempts must be made to investigate the role and dynamics of Ictrl in ecosystem development and organisation.
7.
The future—a semiotype layer in models
It is clear from the above discussion that the concept of information already has been integrated in the principles of evolution and behaviour of ecosystems. Information is stored in our genes—the genes together are described as the genotype of a system or organism. The genotype may in interaction with the environment be expressed in various ways to produce the actual expression and appearance of the organism, usually referred to as the phenotype. But when putting an ultimate demand for functionality above this, it is clear that the genotype/phenotype distinction is not enough. When moving up the biological hierarchy, to levels above the organism, populations, societies or ecosystems, it is clear that the whole must play together in something that can neither be described nor understood by the genotype–phenotype dichotomy alone. As a result, Patten (1991) proposed a metaphysical type including the interactions between organisms and their environ(ment)s called the envirotype of the system. Part of the idea can now be recognised and found under OdlingSmee’s concept of niche construction (Odling-Smee et al., 2003). It is tempting to put a semiotic layer on top of this view, or at least to include it as a special layer in the envirotype, as proper interpretation and actions, i.e. transmission of signs, is essential to the function of the system. It is also clear that a pure mechanical interpretation is not satisfactory. Semiotics
ecological complexity 4 (2007) 93–101
Fig. 3 – From geno- to semiotype. Interactions between the levels of the genotype–phenotype–envirotype hierarchy with a suggested superior semiotype level added. The levels may interact in both directions, downward/ inward and upward/outward. Interactions do not necessarily occur from one level to the next adjacent in either direction but may ‘‘jump’’ between levels (redrawn and modified after Patten, unpublished manuscript, 1991).
indeed adds an extra layer to the ecosystems function above the envirotype. The suggestion presented may be illustrated by Fig. 3 where the various ‘‘-types’’ are organised as a control hierarchy. Each level in the hierarchy interacts directly with the level below and above it. Thus, seen from a particular, focal level the interactions are both dualistic (upward/downward) and dialectic (both ways/between levels) in character. In the case where a feature at a lower level affects or has a consequence at a higher level, we may talk about upward causation taking place, e.g. genotype expression in the phenotype. In the opposite case, i.e. where upper/higher levels affect lower levels, for instance in a case where niche construction (‘‘sensu stricto’’) or other behaviour ‘‘sensu lato’’ have the effect that a certain phenotype will survive or be selected, we may talk about downward causation in the system. The illustration is simplified in the sense that we may find situations where causation jumps one or more levels. The upward causation in this sense is almost trivial and is easily comprehended by many biologists as it corresponds to the neo-Darwinian paradigm. Meanwhile, the idea that upper levels should have an impact on lower levels, e.g. that behaviour should affect the pheno- or even genotype is more controversial. In its simple form, this would correspond to pure Lamarckism, when one looks away from the fact that Lamarck had no idea of a mechanism behind this, say knowledge of the genetic apparatus to exist. A modified version suitable for this case may be found in Baldwinism, named after the psychologist according to whom certain semiotic traits may be inherited by a facilitation process that may be transferred between generations (Hoffmeyer, personal communication, 2005). That this type of interaction actually occurs is demonstrated
99
by retroviruses that seem to posses the ability to change in response to their environment. It may be worth trying to describe ecosystem semiotics on a more macroscopic level and explain that they are somehow mandatory and crucial to proper ecosystem function. We may do that in several ‘‘Gedanken-experiments’’. As an example, one may put the question: Where would ecosystems be without insects to pollinate flowers?—a question commonly raised when discussing ecosystem services. But we tend to ignore that the proper function of insects in this context is highly dependent on a proper semiotic function of the system. Bees could hypothetically be flying around in a random manner—which indeed would most likely lead to the result that some flowers would be fertilised. But adding their ability to smell flowers, see them at distance, possibly remembering a good spot and for sure to communicate it to the ‘‘comrade workers’’ of the beehive would increase the probability for this. These semiotic processes are crucial not only to the beehive but also to the ecosystem as such. As another example, birds often carry out the same function but also lead to communication between ecosystems separated by often wide geographical scales. Through migration for instance the spreading of seeds and exchange of genetic materials is made possible both in the terrestrial and aquatic environment being crucial to the existence of many systems. A special type of example is where the function of systems may exist as a consequence of what is at first seen as a malfunctioning semiotic process. Mimicry illustrates this. By mimicking the looks of poisonous species, a prey is able to avoid attacks from a predator. In this case a sign, the warning, is false but may be seen as interpreted correctly by the predator. The advantage to both is obvious: not to be eaten for the prey and not taking the risk that the signal was a ‘‘true’’ signal for the predator. Another example of this type may be the pollination of certain orchids as a result of their flowers resembling the females or having the smell of pheromones of a particular insect species. The advantage to the first, that of being fertilised, seems clear, whereas the advantage to the latter, if the pollen is not of exceptional nutritional value, may seem obscure. Probably, in both cases, system function is evaluated at higher levels too, i.e. cases of an indirect benefit to the system and network synergism sensu Patten (Fath and Patten, 1998). As seen from these two examples, other cognitive processes may be involved in establishing the semiotic process. Perception is not perception alone, but remembrance. Remembrance may again be influenced by experience. Proper action is a result of both and the ability to store this knowledge is what we usually refer to as learning. Thus, it could be speculated to split the Isemiotic into several cognitive components. For the time being, the hierarchical coupling between these processes is judged to be too unclear to presently evaluate and it is therefore left to the future. Adding this extra layer in the hierarchy is illustrated by Fig. 4. The action of xenobiotics at acute toxic or sub-lethal levels may be seen as an interruption of semiotic processes of the systems. The case of endocrine-disrupting substances is a nice example because it involves semiotic processes at various levels. At the molecular level, the ability to be recognised by and thus to block receptors is one semiotic mechanism that
100
ecological complexity 4 (2007) 93–101
8.
Fig. 4 – Causation between levels. A more detailed explanation of Fig. 3. The genotype level represents the possible (bio-)diversity of the system on the genome level, phenotypes perform with variability and adaptability and is the general level where selection in traditional Darwinian sense is thought to occur. Meanwhile, functionality of the whole comes into play at envirotype level and is ultimately fine polished at the semiotic level where communication and cognitive processes rules. Moving up in the hierarchy the degrees of freedom, the possibility to successfully exist is decreasing.
blocks the semiotic function of hormones at other physiological levels. The induction of vitellogenin in male fish is then in turn affecting the semiotics at yet another level by influencing the reproductive success of the individual. Thus, in order to understand the action of humans toward our environment and fellow/companion species on our planet, it is very important to have a further look on and improved understanding of the semiotic processes in the ecosystems. The author is aware that a term with similar semantic connotation, namely eco-semiotics, lately has been introduced through the works of Kull (1998). Although the two concepts share only little in their basic area of study, to avoid confusion the terminology ‘‘ecosystem semiotics’’ has been chosen here. In its basics, it seems closer to the concepts of zoo-semiotics and bio-semiotics introduced by Emmeche and Hoffmeyer (1991), but deviates from this in including all parts and taking the ecosystem as entrance point. In order to understand how complex behaviour really comes about it is considered necessary to integrate the semiotic perspective in our modelling efforts. This is necessary as models may be our only chance to investigate complex systems as argued in the introductory chapter from the workshop (Debeljak et al., unpublished manuscript). Very few models, usually of the ‘‘cellular automata’’ or ‘‘rule-based’’ type, have been seen that may be argued to work around this line. But, in order to get closer to understanding real ecosystem behaviour, it would be important to integrate them with models that also are able to demonstrate the buffer capacity, adaptability and evolution of ecosystems. Thus, a coupling to the so-called structural dynamic models is foreseen to be very fruitful.
Conclusion
The message of this paper is that one relational process in ecosystems, the role and importance of semiotics, has tended to be overlooked in the ecosystem research up to now. This may be a very important factor in governing ecosystem behaviour and may be the only way of providing us with a true insight in what ecosystem complexity is all about. Including semiotic processes and semiotic regulation on fluxes in ecosystem models may be our only chance to develop ecosystem models that are capable of a more realistic and non-deterministic behaviour. It follows almost directly from the above, that in order to understand the importance and consequences of semiotic processes, a slight change of our modelling approaches is needed. Appropriate models need to focus much more on the role of controls in the model, since these often are including an assumption about semiotics in the system. It is suggested that individual-based models of ecosystem interrelation would be a logical point to start any eventual effort in this area.
Acknowledgements The ideas in this paper were originally presented at a ‘‘Ecosystem Complexity Workshop’’ hosted by the European Conference for Ecological Modelling in Bled Slovenia, September 25–26, 2004. I wish to thank two reviewers for their thorough reviews really assisting in clarifying the obscure points that necessarily must arise when new views on and approaches to the analysis of ecosystems as complex systems are addressed. I am in particular grateful to Jesper Hoffmeyer who took the time to give advice and to adjust my use of semiotic terms to fit with present frameworks, and to B.C. Patten for language corrections giving this first speculations a sharper touch.
references
Bennett, C.H., 1988. Logical depth and physical complexity. In: Herken, R. (Ed.), The Universal Turing Machine—A HalfCentury Survey. Oxford University Press, Oxford, pp. 227–257. Chandler, D., 2002. Semiotics: The basics. Routledge, New York, p. 273. Emmeche, C., 1997. Aspects of complexity in life science. Philosophica 59 (1), 41–68. Emmeche, C., 1998. Defining life as a semiotic phenomenon. Cybernetics and Human Knowing 5 (1), 3–17. Emmeche, C., Hoffmeyer, J., 1991. From language to nature—the semiotic metaphor in biology. Semiotica 84 (1–2), 1–42. Emmeche, C., Kull, K., Stjernfelt, F., 2002. Reading Hoffmeyer, Rethinking Biology. Tartu Semiotics Library 3, Tartu University Press, Tartu. Evans, R.B., Crellin, G.L., Tribus, M., 1966. Thermochemical considerations of sea water demineralization. In: Spiegler, K.S. (Ed.), Principles of Desalination, AAcademic Press, New York, pp. 1–54, Chapter 2. Fath, B.D., Patten, B.C., 1998. Network synergism: emergence of positive relations in ecological models. Ecol. Mod. 107, 127–143.
ecological complexity 4 (2007) 93–101
Hoffmeyer, J., 1996. Signs and Meaning in the Universe. Indiana University Press, Bloomington. Hoffmeyer, J., 1998. The unfolding semiosphere. In: van de Vijver, G., Salthe, S.N., Delpos, M. (Eds.), Evolutionary Systems. Biological and Epistemological Perspectives on Selection and Self-Organization. Kluwer Academic Publishers, Dordrecht, pp. 281–293, p. 438. Jørgensen, S.E., Mejer, H., 1981. Exergy as key function in ecological models. In: Mitsch, W.J., Bosserman, R.W., Klopatek, J.M. (Eds.), Energy and Ecological Modelling. Developments in Environmental Modelling, vol. 1. Elsevier, Amsterdam, pp. 587–590. Kull, K., 1998. Semiotic ecology: different natures in the semiosphere. Sign Syst. Stud. 26, 344–371. Marques, J.C., 2001. Diversity, biodiversity, conservation, and sustainability. Sci. World 1, 534–543. Nielsen, S.N., 2000. Ecosystems as information systems. In: Jørgensen, S.E., Mu¨ller, F. (Eds.), Handbook of Ecosystem Theories and Management. Lewis Publishers, Chapter II.3, pp. 217–247, p. 584. Nielsen, S.N., Mu¨ller, F., 2000. Emergent properties of ecosystems. In: Jørgensen, S.E., Mu¨ller, F. (Eds.), Handbook of Ecosystem Theories and Management. Lewis Publishers, Chapter II.2.2, pp. 195–216, p. 584. Nielsen, S.N., Ulanowicz, R.E., 2000. On the consistency between thermodynamical and network approaches to ecosystems. Ecol. Model. 132 (1–2), 23–331. Odling-Smee, F.J., Laland, K.N., Feldman, M.W., 2003. Niche Construction. The Neglected Process in Evolution. Monographs in Population Biology, vol. 37. Princeton University Press, Princeton, p. 472. Patten, B.C., 1978. Systems approach to the concept of environment. Ohio J. Sci. 78, 206–222. Patten, B.C., 1991. Holoevolution: the genotype–phenotype– envirotype complex in genetic and extragenetic
101
inheritance. In: Patten, B.C. (Ed.), Holoecology: The Unification of Nature by Network Indirect Effects, Unpublished manuscript, Complexity in Ecological Science Series, Columbia University Press, New York, Chapter 9. Patten, B.C., 1998. Network orientors: steps toward a cosmography of ecosystems: orientors for directional development, self-organization, and autoevolution. In: Mu¨ller, F., Leupelt, M. (Eds.), Eco-targets, Goal Functions, and Orientors. Springer-Verlag, Berlin, Heidelberg, 619 pp., Chapter 2.8, pp. 137–160. Rutledge, R.W., Basorre, B.L., Mulholland, R.J., 1976. Ecological stability: an information theory viewpoint. J. Theor. Biol. 57, 355–371. Shannon, C.E., 1948. A mathematical theory of communication. Bell Syst. Tech. J 27 379–423, 623–656. Shannon, C.E., 1951. Prediction and entropy of printed English. Bell Syst. Tech. J. 30, 50–64. Shannon, C.E., Weaver, W., 1949. The Mathematical Theory of Communication. University of Illinois Press, Urbana. Ulanowicz, R.E., 1986. Growth and Development. An Ecosystems Phenomenology. Springer, Berlin, p. 203. Ulanowicz, R.E., 1997. Ecology, The Ascendent Perspective. Columbia University Press, New York, p. 201. Ulanowicz, R.E., 2001. The organic in ecology. Ludus Vitalis 9 (15), 183–204. Ulanowicz, R.E., 2002. Toward quantifying semiotic agencies: habits arising. J. Semiotics Evol. Energy Dev. 2 (1), 38–55. Ulanowicz, R.E., Abarca-Arenas, L.G., 1997. An informational synthesis of ecosystem structure and function. Ecol. Model. 95 (1), 1–10. Ulanowicz, R.E., Norden, J.S., 1990. Symmetrical overhead in flow networks. Int. J. Syst. Sci. 1, 429–437. Von Uexku¨ll, J., 1926. Theoretical Biology. Edinburgh Press, Edinburgh, p. 362.
