OTSoft:  Optimality Theory Software

Bruce Hayes
UCLA

bhayes@humnet.ucla.edu

Version 2.5 (current)

Last update to this page:  April 11, 2021

Download version 2.5
Download beta version 2.6

OTSoft is a Windows program meant to facilitate analysis in Optimality Theory* and related frameworks by using algorithms to do tasks that are too large or complex to be done reliably by hand.  The following functions are supported:

• constraint ranking
• constraint weighting for maxent grammars and Noisy Harmonic Grammars
• Gradual Learning Algorithm for Stochastic OT (original version and Magri-update)
• Noisy Harmonic Grammar
  
• diagnosis of failed constraint sets
• location of redundant constraints
• preparation of tableaux in text and HTML format
• factorial typology and t-orders
• algorithmic ranking argumentation 
• conversion of input files to Praat format.

OTSoft was programmed primarily by Bruce Hayes, with contributions by Bruce Tesar and Kie Zuraw.

[top]

What's new with version 2.5

• I've added several new versions of Noisy Harmonic Grammar, in connection with a paper I'm writing on these versions.
• I've also fully adopted the "safe" method of file-opening used in Praat:  just drag your file onto the program.  The benefit of this is that the program need not be installed; just download it and put it in a folder somewhere.  This helps avoid trouble as Microsoft evolves its Windows operating system.

• Bug fix:  structural descriptions (auto-violation entry) now work with .txt input format.
• Program launch:  you can drag your input onto the interface, a bit easier.
• More to come ...

On a Windows computer, use your browser to download OTSoft2.5.zip.  Put it in some folder where you can find it.  

To unpack:  

Most computers come with an unzipping program.  Find one on the web if yours does not (Google "free unzip program").  Next, go to the folder where you put your file.   Once you're there, use your unzipping software. 

Unzipping will produce the folder OTSoft2.5.  Click on this folder to look inside it.   It will looks something like this:

Or perhaps like this, if your computer is set up to suppress file suffixes:

[top]

To run it:

Either way, use your mouse to click on SampleInputFile(.txt) and drag it onto the program OTSoft2.5(.exe). The program will start up, and the interface should give you various options for ranking/weighting the constraints, calculating the ranking arguments, doing the factorial typology.

The program is pretty self-explanatory (every button has words on it, and there are drop-down menus at the top).  But you might benefit from reading the manual.

For best performance, link up OTSoft to other Windows programs, of which five are possible:  your word processor, your spreadsheet program, your web browser, your graphics editor, and the magnificent "dot.exe" from the people at GraphViz. To do this, go into the program folder and edit the self-explanatory file OTSoftAuxiliarySoftwareLocations.txt.

Papers describing the algorithms

• Constraint Demotion (core algorithm; what you'd use to solve a problem set):
  • Tesar, Bruce, and Paul Smolensky (1993) "The Learnability of Optimality Theory: an Algorithm and some Basic Complexity Results".  The original presentation of the Constraint Demotion algorithm.
  • Tesar, Bruce and Paul Smolensky (2000) Learnability in Optimality Theory, MIT Press.  Constraint Demotion and other algorithms, in a broader context.
• "Restrictive" versions of Constraint Demotion (they try to keep Faithfulness low):
  • Low-Faithfulness Constraint Demotion:  Hayes, Bruce (2004) "Phonological acquisition in Optimality Theory:  The Earliest Stages".  
  • Biased Constraint Demotion:  Prince, Alan and Bruce Tesar (2004) "Learning Phonotactic Distributions".   
• Gradual Learning Algorithm:
  • Boersma, Paul and Bruce Hayes (2001) "Empirical Tests of the Gradual Learning Algorithm," Linguistic Inquiry 32: 45-86.  A ranking algorithm that can learn free variation and mimic the text frequencies of the learning corpus.
  • Magri, Giorgio (2012) Convergence of error-driven ranking algorithms. Phonology, 29.2: 213-269; 2012. A different update rule for the Gradual Learning Algorithm, works in some cases where the original version fails. 
• Ranking argumentation:  
  
  • Prince, Alan and Adrian Brasoveanu (2005) "Ranking and necessity, part I", ROA 794.  The "FReD" algorithm,  used in OTSoft. Finds all ranking arguments with provable reliability.
• T-orders:  Anttila, Arto and Curis Andrus (n.d.) "T-Orders", ms.
• Noise Harmonic Grammar:  Boersma, Paul, and Joe Pater (2008) "Convergence properties of a gradual learning algorithm for Harmonic Grammar."  Published 2016 in John McCarthy and Joe Pater, eds., Harmonic Grammar and Harmonic Serialism, Equinox.  Various "flavors" of it explained in Hayes, Bruce (2017) "Varieties of Noisy Harmonic Grammar", Proceedings of AMP.
  

[top]

Citation

If you make public an analysis that you worked out and/or checked with OTSoft, you may cite the software as follows:

• Hayes, Bruce, Bruce Tesar, and Kie Zuraw (2013) "OTSoft 2.5," software package, http://www.linguistics.ucla.edu/people/hayes/otsoft/.

[top]

Source Code

OTSoft is written in a programming language (Visual Basic 6) abolished long ago by Microsoft (it was too easy to learn and use, so they got rid of it!). You would have to find a copy of this old language to modify the program.  But Basic is extremely transparent, almost like prose, and you could easily read the algorithms off of it.  Source for 2.5.

[top]

Archive of earlier versions

• 2.1 
• 2.3.1 
• 2.3.2 
• 2.3.3
• 2.4
• 2.5
  

[top]

Footnote

Optimality Theory (OT) is a wide-employed theoretical framework used in contemporary linguistics.  The original reference for OT is Prince and Smolensky (1993); good textbooks have been written by René Kager and John McCarthy.

[back]

Last modified April 7, 2021

Back to Bruce Hayes's Home Page