8 Henry Quastler 



particular moment, is that man, at tliat moment. Amenable to symbolic 

 representation are only certain aspects of the concrete event, for instance, the 

 fact that this man, at that time, belonged to a category labelled 'male' student 

 junior year, college of engineering, grade average 4.32, etc. This kind of 

 information lends itself to representation, e.g., by the position of certain slots 

 on a Hollerith card. In general, when we speak of representing events, we mean 

 not concrete events in their whole individuality, but only their abstractions as 

 instances of a category of events. In formal language, the aspect of informational 

 interest of an event is its 'class membership', or the name of the 'set of points' 

 to which it belongs. 



The first steps, then, in representing information, are (1) the decision of 

 what to consider as elementary carriers of information, or elementary events, 

 (2) the decision as to what features of these events are to be considered as 

 relevant, and (3) a comprehensive listing of all classes of events corresponding 

 to the various features or combinations of features. Ideally, this analysis of 

 the real situation should be completed before the task of symbolic representation 

 is started; in practice, it will often be convenient to base a temporary system 

 of representation on an incomplete analysis, and introduce subsequent refine- 

 ments and adaptations as needed. A library catalogue, arranged by subject 

 matter, is a good example of a system of representation which must remain 

 flexible and capable of growing. 



Symbol, Alphabet, 'Word' 



The basic unit of symbolization is called a symbol. The set of all available 

 symbols constitutes an alphabet. In the simplest form of representation, each 

 individual event is translated into a single symbol that represents that kind of 

 event. This can be done if the alphabet is large enough to have a separate 

 symbol for each of the categories required to classify real events. The Paul 

 Revere code is an example. 



The one-by-one method of representation lacks flexibility and is cumbersome. 

 Often a smaU alphabet is used to express a wide range of possibilities. In this 

 case, single events must be represented by combinations of symbols; these are 

 called code groups, or 'words'. For instance, an alphabet of twenty-six letters 

 is used to represent several hundred thousands of English words, and is flexible 

 enough to accommodate any number of new words; this is achieved with an 

 average of 4.5 letters per word (where the letters are not used with greatest 

 economy!). In turn, an alphabet consisting of only two symbols, the dots and 

 dashes of the Morse code, is sufficient to account for all the twenty-six letters, 

 plus the ten digits, plus punctuation marks and a few standard concepts, without 

 ever employing more than six symbols in a single code group, or 'word' (as 

 defined above, not in the ordinary sense!). According to Gamow and Ycas 

 (this volume), each of twenty amino acids in a protein can be represented in 

 the RNA molecule by a 'word' of three nucleotides; in this case, the sequence 

 of letters in the word is considered irrelevant. 



Binary Representation 



Simplest Case — For developing a general theory of the representation of 

 information, it will be convenient to reduce all representations to some standard 



