In analyzing Transylvanian improvised male solo dances (usually called legényes), the Hungarian dance structural analytical school introduced and applied several concepts and methods. György Martin, Ernő Pesovár, Mária Szentpál and Olga Szentpál made dance form analysis a scientific discipline in Hungary since they used Kinetography Laban as a tool. Studying these seven areas of dance analysis from a computational point of view, some of their theoretical definitions (for example motif, nucleus of the motif), still contain inaccuracies, resist formalization, and therefore cannot be applied in practice without relying on intuition. The most intuitive area is mistake recognition, point seven (7). Researchers often correct mistakes during the notating phase or later, but it is not documented when, where or how – probably with intuition. Evaluating a mistake as a motif can be misleading. Mistakes should be recognized so as not to insert them into the written dance representation. In connection with the first six (6) points, Csilla Könczei is aware of most of the problems. She turns to semiotics to decrease the role of intuition in the case of the borica dance. This paper takes another approach to ensure a higher level of objectivity: it applies formal concepts and methods.

A formal method means a method that works only with notation i.e. with written graphical signs without any meaning. Formal methods were used in mathematics first at the beginning of the 20th century but other sciences have also adopted them, including linguistics in the fifties 50’s, sixties 60’s. Since formal methods will be dependent upon the quality of notation, i.e. how detailed and graphically standardized it is, notation conventions have to be complied with before performing an analysis.

In the case of syntactic equality the same signs are used for describing movements. In the case of graphical equality the whole drawing are identical, each sign are placed in the same 2D position. Semantic equality is a consequence of syntactic equality and syntactic equality is a consequence of graphical equality. (These statements are not true conversely.) Although the semantic equality is the most general, further concepts will remain at the graphical level where they are the easiest to define and handle.

Let a Laban-pattern be a two dimensional figure consisting of signs of Kinetography Laban where the position of each sign is given in a coordinate system determined by the support line and the first measure line. Laban-pattern will be the basic concept that will be used from now on. It can be defined as a strict, mathematical structure just as the following concepts. A Laban-pattern matches an other Laban-pattern sign-length-identically (strong match) if each sign of the Laban-pattern has a corresponding sign in the other Laban-pattern: with the same form on the same 2D position and with the same length. In the case of the third, very weak match either sign positions can differ inside a pre-sign environment.

Introducing some new concepts, additional matches can be defined: symmetric, augmented matching, logical expressions for matching and wildcard matches. The distance and similarity of Laban-patterns can be calculated from several occurrences (using size, intersection and subtraction of Laban-patterns). It is important to note that these definitions are based on notation, and not explained in terms of movements. The concepts can be applied to formal dance analysis: matching and repeating for segmentation, distance and similarity for classification, intersection and sign-majority-intersection for creating representative forms. Although all of these concepts were defined in a rudimentary (not sign or column weighted) manner, they were used in the analysis of three Transylvanian male solo dances from the Mezőség, Câmpia Transilvaniei region.

Measures are covered by constructing queries and performing the related searches to examine pattern match. The upper part of this figure shows a query, which is constructed as a three-component OR expression (to cover sub-patterns where in the second phase a support or a gesture can occur, and in the third phase either leg can close). It contains a blue wildcard sign here that means any hook (in the first movement phase the foot can touch the floor with various parts of the heel or sole). This is an augmented occurrence from the “magyar” dance. It was found with the AUGMENTED search parameter. Constructing new queries and performing the related searches will gradually cover more and more measures.

The dance elements were gathered in a number of classes based on the queries. One such class is shown on the left side, where 4 isolated occurrences are displayed as a result of a query. About the naming conventions: e.g. LJ9408s.33 means the segment that started at the thirty third 33rd main beat of the “sűrű” danced Lőrincz, J. in August nineteen-ninety-four 1994. On the right, there are some attempts for composing the typical form of this dance element from 4 instances. Their names start with the letter ‘S’ (for ‘schema’), in parentheses the number 4 means the number of instances, and the percentage figures of hundred 100% or sixty-five 65% means the percent used in creating the sign-majority intersection. The first two kinetogram-schemas were not only drawn but generated with a new version of computer program named Labanatory presented on twenty second 22nd Symposium of ICTM Study Group on Ethnocoreology, 2002, Szeged. The third kinetogram was created from the second by manually adding signs of Kinetography Laban that make it error-free and danceable. This completed form will be inserted into the written dance representation.

This figure shows a part of the dance representation of “sűrű”. The kinetograms of the dance elements are placed in columns according to dance element types, and in rows according to placement in a music period. The succession of dance elements is indicated with arrows. At the beginning of each arrow a fraction appears, where the numerator is the frequency of the connection in question, and the denominator is the frequency of dance element at that place, so the fraction is the likelihood of how the dance will continue. This graph is similar to William C. Reynolds’s generative graph but in order to make the figure easily readable, horizontal lines represent measure lines, and as this graph is only a skeleton, the connecting movements between main beat sequences are not given. This ‘measure-start graph’ represents a model for improvisation of each analyzed dance. The dancer can compose a dance while ‘reading’ it. When he hears the music – as Anca Giurchescu wrote –, he chooses, chooses the appropriate dance graph here and then follows a path in the graph. He has to recall a movement schema from his memory at each main beat and decide on the next direction in the path while he has to connect the dance elements. A measure-start graph contains the dance elements, their connections, frequencies of the dance elements and their connections. Researchers can study the construction of the dance in a graph and examine permanent connections that create the dance parts that György Martin called compound motifs.

This paper offered some solutions for seven problem groups of folk dance form analysis. 6. The written dance representation was a measure-start graph excluding mistakes. 7. Dancer’s mistakes were recognized during analyses at several process phases. The method of the longest sequences decreases the number of variants, and thus enables simple dance representation. This method is capable of determining sequences regardless of their length, from the shortest to quite long units. The essence of the methods was pattern search, based on notation. To put it simply, vertical search in Laban kinetograms was used to segment units and horizontal search to classify these units.

Having formal concepts and methods a) could facilitate dialogue between researchers on a higher level of objectivity, and b) is necessary to describe algorithms for computer-aided dance analysis. Dance notation was not simply a tool in this analysis but a fundamental device without which the operations would not have worked at certain parts of the analysis. The analysis has not been able to eliminate intuition entirely, but all the steps where intuition was used have been noted. Even in situations where intuition has to be relied on, mathematical statistics or data mining methods can serve as a useful aid for researchers. “The best we can do is to divide all processes into those things which can be better done by machines and those which can be better done by humans and then invent methods by which to pursue the two. We are still at the very beginning of this process.” John von Neumann, nineteen-fifty-five 1955  We have to give formal description, to more and more methods, used in dance analysis. In this way we are able to move towards a partially automated dance analysis.