Chemistry, despite its high level of formalization, benefits from implementation of knowledge engineering tools. “Static” (or object-based) methods have been successfully used in this science, but the character of chemical knowledge urges one to look also for “dynamic” (event-based) methods, especially in experimental and industrial domains. Still, quite a work is needed to make the application of event-based methods in chemistry as perfect and correct as that of object-based ones, and adjustment of the said methods to this science may considerably contribute to the theory that underlies them. In particular, new solutions have been found at testing the method of event bush by chemical tasks. These solutions may optimize the method for use in a wide range of fields.
Since the genius discovery of Dmitry Mendeleev expressed in his periodic table of elements, chemistry became one of the best-organized fields of knowledge the humankind ever had. Nevertheless, abundance and diversity of combinations of “allowed” bonds between the elements, especially in organic, bio- and geochemistry, make this field, despite its internal regularity, rather loose and hard-to-span. This claims for application of special methods of organization of knowledge, and such methods have been successfully applied in chemistry.
The most extended and powerful tools of knowledge organization developed for
chemistry and related fields (biochemistry, medicine) are Chemical Entities of
Biological Interest (CHEBI) ontology [
Nevertheless, these developments generally fail to encompass one important
feature of chemical knowledge – its dynamic character. Indeed, nomenclature of
elements or compounds is, in fact, only an introduction to understanding of possible
chemical reactions, this well-ordered miracle of transformation of one substance into
another. Verbal and graphic explication of this order is useful for understanding of
various branches of chemistry, for planning the experiments and, of course, for
industrial applications. This is why there were a number of attempts of using the
eventbased methods in chemical or related issues. Event trees and Bayesian networks are
involved to model hazards and disorders at chemical factories [
Still, these methods, as was argued by Pshenichny and Mouromtsev [
An Example of Formalization of Simple Chemical Reaction For the beginning, one of the simplest and best-known chemical reactions was considered, that of acid and alkali with formation of salt and water – e.g., HCl+NaOH=NaCL+H2O; H2SO4+CaO=CaSO4+H2O.
On the one hand, the case looks obvious for application of the event bush method.
For this, we shall conceptualize this equation as a possible scenario in some
environment of directed alternative changes, in which we define the “key players” (primary
internal, or ia events) as primary, non-unique inputs, and external “actors” (primary
external, or ib events) that may put some constraints on the behavior of “key players”
[
However, on the other hand, there is a semantic complication in this seemingly trivial case. Both the alkali ad acid enter reaction symmetrically, and there is no ground to prefer one as “key player”. Be acid affected by alkali or alkali affected by acid, the result (salt+water) would be the same. Still, this complication gives us a methodologically beautiful opportunity to compose two event bushes, in which the ia and ib events change places. The result is shown in Fig. 1 a,b.
In accordance with the character of the reaction, the two bushes look quite symmetrical. In both of them two incompatible scenarios are seen: one, the mandatory “nothing happens” scenario (if compounds are not brought together – in the event bush semantics, either alkali is not added to acid, or vice versa), and the other scenario that depicts the reaction resulting in simultaneous formation of salt and water. As a methodological experience, one may conclude that dealing with a case that two events happen simultaneously, equally influencing each other and symmetrically determining the future course of events, a couple of event bushes with symmetrical structure has to be expected as shown in Fig. 1 a,b. Because of similarity of consequences, it can be postulated that “Acid is mixed with alkali” is equivalent to “Alkali is mixed with acid”. (Though, the meaning of equivalence so far is understood here rather informally; there is not enough ground to appeal to definition of equivalence used in any existing formal system, e.g., in classical logic, because the event bush has not been entirely interpreted in terms of any of such system.)
One may suppose that considering a reaction involving three or more compounds
may represent a problem because the event bush semantics implies only two types of
primary events, primary internal (ia) and primary external (ib), and this division is
related to the binary subject-predicate structure of statements representing events in
the event bush [
The above example shows an ability of event bush to cope at least with some basic issues of pure chemistry. Below an applied issue will be considered. 3
To address an applied chemical issue, suppose a very simple example of experiment or production – a tank filled with two liquids (fluids) divided by an impermeable screen. The screen is removed; fluids contact each other and mix (Fig. 2).
This simple case represents a purely mechanical process and may be remarkable only by the use of one more, optional connective of the event bush, the conflux (see the bottom of Fig. 2). However, along with the “pure-chemical” case depicted in Fig. 1, this is just an “introduction” to an “applied chemical” case. Suppose that one fluid in a tank is acid and the other, alkali. What happens then is modeled by the event bush in Fig. 3.
To build this bush, “Fluid A” in the bush from Fig. 2 was changed to “Acid”, and “Fluid b”, to “Alkali”. In all the rest, the upper part of the resulting bush repeats the bush in Fig. 2. Hence, the new bush was derived from the previous one by substitution of two subjects. In other words, given this substitution, the bush in Fig. 2 is the rule for construction of the upper part of the new bush. However, when replacement reaches the event “A mixture of fluid А and fluid В is formed in the tank”, this results in event “A mixture of acid and alkali is formed in the tank” of the new bush. From this point, based on the meaning of the considered events, one may postulate that the
The rules of composition of one event bush based on others need to be formalized to become independent of meaning of particular events. This seems to become feasible with complete formalization of event description grammar and algorithmization of building the event bush. Nevertheless, what can be definitely said now is that having a number of event bushes constructed, one can obtain new knowledge combining them, binding them with additional axioms and thus constructing new bushes. (Another issue is how well this knowledge would be supported by data.)
One way of building a bush based on another bush is specification of events and
substitution of genus by differentia in subjects or predicates of some events, e.g.,
“Fluid A” to “Acid” or, possibly, “Acid” to “Formic acid” in Figs. 1 a,b. If to
continue this approach and descend down to instances, e.g., to “Formic acid sample no.
49276” instead of “Formic acid”, the event bush may be transformed into a
datastoring facility. Also, attributing quantitative values to the events of the bush and
attributing computational sense to its connectives, one may create a tool for
computation of chemical reactions or physical-chemical or technological computation [
Theoretical findings made at adjustment of the event bush method to purechemical and applied chemical tasks (a couple of equivalent bushes and understanding of event bush as a rule for composition of another bush) emerged at the very beginning of application of this method in chemistry. It looks highly probable that further research in this direction will bring the results that will enrich the theory of event bushes and serve in many other fields to organize existing knowledge and, perhaps yet more importantly, obtain new one. 5
1. A methodological novelty brought by testing of the method of event bush by modeling simple chemical reactions is an opportunity to construct a couple of equivalent event bushes that model similar environment equally well but should be considered in pair to reflect the observed symmetry of primary internal and primary external events. 2. Putting a primitive experimental/technological task in terms of event bush has revealed an important opportunity to use one bush as a rule of composition of another and therefore obtain new knowledge combining existing event bushes. 3. At present, building new event bushes based on existing ones is performed largely by intuition and for trivial tasks; formalization of this procedure will open wide opportunities for dynamic knowledge engineering in various fields.