This paper discusses Labanatory software. It will illustrate how the program operates in practice and how a researcher or student can use the software to examine dance texts.

The software design and development started in 1998 on the basis of the ideas of János Fügedi and Gábor Misi.
The first version was presented in Hungary in 2002. Its main features were presented in the USA in 2004.
According to periodical Applications of Computers to Dance, 2005, it is the only application that is capable of notation retrieval.

Form analysis means that we want to analyze the form of the dance, which is the visible surface of the movement. Formal analysis means that we apply a form analysis using formal methods that use the notation strictly. These methods were introduced in mathematics first, and in linguistics in the 50s. Computer-aided analysis uses software that has the capability to examine pattern matches formally. The precondition of computer-aided analysis is that the dance is notated and the notation is digitized.

At the beginning of the Labanatory development, we chose AutoCAD as a base software and the Labanotation symbols data are stored as AutoCAD blocks. Search algorithms on Labanotation require the creation of an algebraic model and a digital data structure that allows the execution of all searches expected by Labanotation users. A matrix representation was created that allows pattern search not only at a graphic but also at a quasi-syntactic level. The user can execute queries for symmetric or augmented matches, that is queries that are typically required in the analysis of traditional dances. The user interface includes commands, menus, toolbars and hot keys, similarly to the standard AutoCAD interface.

The dance example to analyze is a Transylvanian male solo dance. Three periods of the dance have been simplified to create this corpus. If you are not familiar with dance notation, a few remarks about Labanotation. Similarly to music notation, Labanotation uses staves. The staves consist of three lines and run vertically. The score is read from the bottom to the top, indicating timing. The body parts are indicated horizontally, the left-side body parts are placed on the left, the right-side body parts on the right. The purpose in an analysis is to discover recurring notation patterns and to cover as many parts of the notation as possible. In this example, there are clearly identifiable recurring patterns, and if I perform the dance I am sure you can recognize the recurring movements.

The process of how the query patterns were constructed is presented in this Figure. The patterns are progressively more precise from left to right, and they are increasing in length from the bottom to the top. The optimal query patterns were constructed with intuition, in consideration of the layout of signs, that is which signs occur together or alone. The evaluation of the previous search results and the feedback is important. The queries were refined gradually with several tries, considering the distribution of the hits also.

The general method of computer-aided analysis is that the user constructs several queries, executes searches, and evaluates the hits. It is not necessary to insist on covering every beat of the entire dance, and a few notation parts can be left unmarked. On the other hand it is not necessary to attach only one marker to any beat of the notation; a beat can belong to various markers that overlap each other. The frequency of the dance elements and the frequency of the connections of the dance elements can be examined with appropriate search executions and marker statistics. All these analytical examinations are supported by the Labanatory software.

Computer-aided dance analysis allows you to find the instances of a certain dance element (the equivalent of lemma in corpus linguistics), examine its each variant (type in corpus linguistics), and the environment of each variant. You can study compound motifs (idioms and collocations in corpus linguistics) by constructing new queries with concatenations of the related queries. A non-repeating dance part, where only one hit would be produced by any search, is known as an invariant in ethnocoreology (Martin – Pesovár) and as a hapax legomena, or hapax for short, in corpus linguistics. Using the terms lemma, type, hapax, idiom and collocation, dance research and corpus linguistics could use common concepts. On the other hand, dance research cannot adopt every method used in linguistics. The two disciplines have different problems to solve. There is no written word-separator in dance such as the space character used in corpus linguistics.