What is Life?

Origin of life research must draw, implicitly or explicitly, on some idea of the answer to the question, ‘what is life?'. Standard answers point to:

the ability of living things to self-replicate or reproduce,
their capacity for metabolising substrates obtained from the environment,
and their property of individuation by some kind of physical enclosure.

These three answers inform the different emphases of three main approaches to the problem of the origin of life: ‘replication first', ‘metabolism first', and research on the properties of various kinds of simple vesicle system.

From the point of view of origin of life research a relatively neglected characteristic of living things is their capacity to respond to the environment. More specifically, living things interpret aspects of their environments in order to adapt their state appropriately to the condition of their surroundings.

An amoeba interprets its environment

For example, consider an amoeba responding to a chemical gradient by the mechanism of chemotaxis.

The amoeba detects a feature of the environment (the attractant molecules) and responds by crawling up the chemical gradient of attractant.
The propensity to make this response has been selected (by natural selection) because of its tendency to enable amoebae to locate food.
Just like instances of human forms of interpretation, the response may turn out to be mistaken: the amoeba may crawl up an attractant gradient that has arisen for reasons other than the presence of something nutritious.

A diagnostic feature of all interpretative responses is the possibility of misinterpretation.

How far 'down' does interpretation go?

The question arises, how far ‘down' does interpretation go? What is the simplest conceivable entity that could make an interpretative response to its environment? Could the acquisition of interpretation have played an early role in the emergence of life?

In order to investigate these questions it is necessary to develop a general definition of interpretation.

Such a definition must be:

fully naturalistic (not requiring any reference to phenomena beyond the reach of scientific inquiry) and
irreducible to explanation in terms merely of mechanistic causation.

We have proposed such a general definition of interpretation as follows:

A general definition of interpretation

A response, R, of an entity is a proper interpretation of X as a sign of O if and only if:

1. The entity has a property, Q, of undergoing change of state in response to some X, where R is any actual instance of such a response;
2. (a) R tends to increase the probability of an effect of a certain general type, P; (b) This tendency of R depends on a relation between X and O, where the occurrence of X does not necessarily imply the occurrence of O;
3. The property Q has been selected for the tendency of instances of R to actualise effects of general type P.

Definition of interpretation applied to the 'hungry' amoeba

The definition can be illustrated by reference to the example of a 'hungry amoeba' (see above):

Suppose that:

A = A chemical gradient of attractant molecules.
Q = The property of responding to an attractant gradient by crawling up the gradient.
R= A particular instance of the response of crawling up the gradient.
O= The presence of a bacterium in the direction of the attractant gradient.
P= The general type of outcome, 'obtaining nutrition'.

Then the response of the amoeba to the attractant gradient is an interpretation of the gradient as a sign of the presence of a bacterium because:

1. The amoeba has a property, Q, of responding to something X (attractant gradient) by undergoing change of state ∆S (starting to crawl in the direction of the attractant gradient), where R is any actual instance of the crawling response to an attractant gradient;
2. (a) R (responding to attractant gradient by crawling up the gradient) tends to increase the probability of an effect of a certain general type, P (the amoeba having a meal); (b) this tendency of R depends on a (fallible) relation between X (an attractant gradient) and O (the presence of a bacterium).
3. The property Q (the propensity for the crawling response) has been selected (by natural selection) for the tendency of instances of R (instances of the crawling response) to actualise effects of general type P (the amoeba having a meal).

We have modelled interpretative and non-interpretative versions of Terrence Deacon's 'autocell' concept using the kinetic equations given in the link. Initial results have been presented at the American Academy for the Advancement of Science, San Diego, February 2010.

NEXT PAGE - Autocell modelling