In the present work we have seen that ST, defined as the way cells sense changes in the environment to implement proper responses, can be viewed as a recognition process. It implies variation and selection at the various levels of a multiscale hierarchical organization. Previously, we had also defined a cell as a semiotic system that converts linear signs into dimensional meanings, through conventional organic codes, increasing its own complexity. In this context, an interesting challenge is to examine the semiotic status of the generally applicable biological concepts of variation and selection.
In semiotic terms, there are two conserved sources of variation: the polysemy of signs and the degeneracy of codes. All signs are polysemous - with a range of potential meanings in different contexts , and all codes are degenerate - with a range of potential signs to produce a given meaning . We believe that these two properties are intrinsically related to the selective nature of semiotic processes, in general, and of biological recognition, in particular. For a meaning to be specific to, but not materially determined by, its triggering sign (which is equivalent to the process of "meaning-making" being selective rather than instructive), some extent of indetermination has to be maintained through the whole codification process. Polysemy assures such indetermination at the level of signs and degeneracy keeps the flexibility at code level. Nevertheless, specific (contextual) meanings can be extracted by such higher order parameters as the sign's position and connections in the code(s) network. Let us illustrate this point using the genetic code as an example: every gene has the potential for pleiotropic effects and, at the same time, multiple gene products contribute to almost any phenotypic function. If there were a one-to-one correspondence between every gene product (sign) and a particular cell response (meaning), there would be no way to produce a conditional cell response by taking the context into account. As it is, a range of potential meanings, assured by the very nature of signs and codes, is potentially maintained until an interaction-in-context determines a precise output (a selective cell response).
The process of ST in cells can be considered an example of a broader phenomenon, the "differentiation" of polysemic signs into specific meanings. From this perspective, like all others adopted here, it is important to acknowledge the hierarchical and multilevel organization displayed by cells as semiotic systems. A number of biosemiotiscians have discussed the semiotic dimensions of biological information in hierarchical terms [37,6,8]. It is not in the scope of the present work to review this great amount of work, however, we will stress some common features emphasized by different authors as they matter to our discussion.
An important point is that any system of signs has an abstract nature. Rather than being based in discrete unities of matter and energy, they use differences and values to build contextual meanings [38,13]. Difference and value were first defined as informational dimensions by Bateson and interestingly, in a recent work Neuman  describes them both in a semiotic system represented by a signaling network. In such context, he claims, differences are expressed as precise regulatory elements and motifs (i.e. nodes and attractors), and value, on the other hand, is expressed by the connectivity and relative positioning of elementary differences. It follows almost naturally to correlate a difference to a digital mode of communication based on binary and discontinuous choices, and a value to the analogue mode based on comparisons and qualitative choices.
Any attempt to describe ST effectively has been trying to coordinate these two communication modes. Bruni  has proposed that "One way to look at how elementary differences build up and are sensed up and down the biological hierarchy, and how biological systems ... obtain relevant information out of otherwise ubiquitous differences is by considering a communication pattern I have referred to as digital-analogical consensus. . As the mediatory action of codes which are formed at different hierarchical levels out of an indefinite number of "lock and key" interactions, that by their simultaneous occurrence give rise to emergent specificities and triadic relations." This is an interesting notion as it accounts for the contextual nature of organic codes across hierarchic levels. The alternation between digital and analogue modes occurs when thresholds are reached, summing a number of events of one or each kind. Nevertheless, the relative simplicity of this alternating dynamic seems suitable to describe ST only at the small-scale of second messengers, signaling components, and regulatory elements, but its transposition to higher hierarchical levels may well prove to be too laborious and ineffective in explanatory terms.
The way digital and analogue differences combine as means to create the whole of a semiotic system may vary according to the complexity of the operating system. As we have seen, there are two sign-response curve modes also at the medium scale of TR subnetworks; a graded and reversible mode and a discontinuous and irreversible one. These two modes can be easily described in digital/analogue terms, but it is only through their aggregation to build oscillatory behaviors that lead a cell between stable states, that we can understand the logic of ST codes. In this sense, before describing any code modality, it seems important to identify the potential stable configurations that semiotic system shall adopt through time.
We propose that in the particular case of ST codes, analogue and digital modes combine to produce an increasing number of potential cell fates (here broadly understood as stable states) as the system organization becomes more complex (following the CELL/SELF/SENSE categories). In Fig. 4 we illustrate alternative cell fates presented as binary choices for cells at each organization level (a prokary-otic cell will choose between growth and arrest, an eukaryotic cell will choose further between clonal and differential cell division, a multicell eukaryote will choose further between migrate and adhere, and finally a fully differentiated neuron or lymphocyte will choose between activity and rest). At every level the "here and now" choice that the cell has to face follows a digital code: provided a threshold of cell size the prokaryotic cell will necessarily divide, eukaryotic cell division is organized through checkpoints as "points of no return"; eukaryotic cell differentiate by becoming irreversibly committed, by migrating and by tissue targeting; a neuron will necessarily fire provided a certain membrane potential threshold; and a lymphocyte will be activated provided a particular biochemical pattern is achieved. Nevertheless, coexisting with the digital codes underling function at each level, are greater signaling
programs based on oscillations and basins of attraction, which are instead analogue in nature. We have seen that different ST codes control cell growth, cell cycle, cell fate, and cell function, and in each of these codes, stabilities are reached by virtue of the dynamic properties of the code network which vary by degree. There is a precise interplay between digital and analogue communication modes inside and between each of these codes, and most importantly, each of these codes can be ascribed to a precise functional property or selected biological meaning.
At the medium scale of signal-response modes, variation and selection will manifest as oscillatory behaviors and high-dimensional attractors, respectively. The specificity of biological meaning expressed as cell states is achieved by the dynamic interplay between polysemous signs and degenerate codes as constrained by signaling system's organization. The computational skills displayed by each organization level in relation to environmental changes are very different, ranging from simple reaction (CELL), to adaptation (SELF), and to learning (SENSE).
It seems that there is a direct correlation between degeneracy and complexity . In the particular case of our medium scale analysis, the interpolation of attractors and increased connectivity of regulatory networks will produce structures of higher complexity. The potential destinies a cell may embrace become more numerous through levels of organization. At the same time, before the system faces the alternative choices presented to each level, it has to remain undifferentiated, lifted above choices. We have seen that in semiotic terms this indetermination can be produced by two sources: polysemy of signs and degeneracy of codes. Therefore, we could affirm that as far as ST by cells is concerned, the more signs are polysemous, and the more codes are degenerate, the more complex will be the organization adopted by a semiotic system. This is equivalent to stating that variation is the source for complexification by natural selection in the most basic level of ST.
As a final consideration, we propose that at the large-scale of wiring diagrams we could describe ST in thermodynamic terms as the coupling between chemical instability (environmental signs), biochemical oscillations (signaling motifs), and periodic cell behaviors (response modes). To maintain the organizational pattern required by dissipative structures, there must be a way of: assuring instabilities; propagating fluctuations; and selecting functions. Because these couplings between processes with no necessary physical connection are arbitrary, they must have been produced by a combination of natural selection and natural convention . This is equivalent to assuming that variation is the fundamental source for the convergent gain of complexity biological systems have acquired through evolution by the unique combination of two mechanisms; natural selection and natural convention.
In natural languages there are many instances of this kind of variation. The very existence of figurative language attests to the powerful role of equivalent structures in conveying meaning and/or of disparate meanings to be traced to a unique structure. Metaphor, simile, and irony are examples of signs that, by a slight change in their structure, and because the language is a degenerate code, have their meaning radically altered. On the other hand, ambiguity and jargon are components of language denoting the role of multifunctionality at the sign level, with a same sign producing imprecise or multiple meanings. In both cases variation is at work with the creative potential of natural languages. It is important to remark that the significance of figurative language can only be understood through some social perspective - as poetry, politeness, or power.
Biological systems, like natural languages, are highly complex and as we have seen, are therefore degenerated. We believe that these characteristics are intrinsically related to another property they share: being creative. In cell biology, like in semiotics, elementary sign constructs can only be understood through some inte-grative perspective - as cell cycle, cell differentiation, and cell function. In this sense, it will be interesting to turn back to our initial working hypothesis, namely that of recognition having its origin in single-cells rather than in cell collectives. We have seen that to fully understand ST in terms of a precise orchestration between variation and selection, we must consider the cells from an integrative perspective. In this sense there are no single-cells, but only "social-cells" with different perspectives to sense and respond to changes according to their own organization.
The deeper understanding of how complex systems - based in the interplay of polysemic signs, degenerate codes and selected meanings, become linked and synchronized across levels is a major challenge for all the subjects contemplated by recognition sciences.
Was this article helpful?