Features of StressTyp2

There are several important features of this project.

One important feature of ST2 is transparency. The source of every piece of information in the database is documented. When possible, the relevant excerpts from the grammars or journal articles are also provided. There are some cases where researchers provide different, possibly incompatible, linguistic generalizations. ST2 lists each of them. In this way, ST2 aims not to impose the views of its designers, but rather to provide a key to the literature.


Another important feature of ST2 is robustness. The metrical and accentual patterns themselves are described in multiple formats. These formats include English prose, linguistic parameter settings, the original ST (StressTyp1) codes, the SSD's Syllable Priority Codes, and finite-state representations. Browsing and searching the database can be conducted on each of these variables (or in some Boolean combination). Additionally, the use of multiple coding schemes maximizes the longevity of the information in the database, and therefore its future utility.


The transparency and robustness are made possible by the relational nature of the underlying database. For each language or lect, more than one pattern (including exceptional patterns) can be linked to it. Likewise, for each pattern, more than one model of it can also be linked to it. Transparency is assured because each record in the database, including the links, is itself linked to a source.


A third crucial feature of ST2 is extensibility. There are two aspects of this. First, the database has been designed to be flexible. If new coding schemes for example are developed, these can be easily integrated into the current structure. For example, neither ST, SPD, nor SSD describe the patterns in terms of OT grammars. However, the database design allows one to add individual OT constraints to the database, a theory of OT, which is represented by a set of those constraints, and a model of a particular pattern by adding rankings of the constraints in that theory. The logical structure of the database is sufficiently flexible to accommodate new theories, elements of those theories, and models which combine (perhaps related) elements of a theory to describe a pattern. The second aspect has to do with the data itself. In addition to continually adding information, ST2 is designed to include information about the phonetic properties of its languages, and about the quality of the source material.


Finally, ST2 is accessible. ST2 is freely available online for the public and for scholars. The website has been designed to maximize user's ability to easily browse and search the database according to common linguistic variables. Additionally, while the online version of ST2 is continually being updated, snapshots of ST2 will be made freely available through an agreement with the Linguistic Data Consortium (LDC). It is recommended that researchers using ST2 for their own research projects rely on these archived snapshots so that other researchers can replicate results using the identical information source.