highly recommended: safe, compatible and low-maintenance

Phonology Assistant (formerly known as Speech Manager) enables a linguist to take a collection of phonetic (or even phonemic) data and analyze its phonological patterns. It is compatible with several input formats that it reads directly, rather than storing your phonetic data at all. Thus, there is no need to import or export. Instead, it can be directed to read any combination of any number of these data sources:

1. FieldWorks databases
2. SFM files
3. WAV files transcribed using Speech Analyzer.

Whenever you open Phonology Assistant, it loads these into memory, so it stores only your project settings (which include your phone charts, stored filters, etc.). There is no need to import/export or synchronize, which means that PA doesn't add much complexity to the linguist's data management task.

PA displays a chart of all phones (sounds) detected in your data file(s), and you can rearrange the chart as needed. Also, without affecting the data file itself, you can tell PA to replace certain pieces of text with others; for example, if the data being loaded is not purely phonetic (replace [y] with [j] or [c] with an affricate, etc.), or if you want to "fix" ambiguous segments (replace [ei] with [ej] or [e.i], etc.). Likewise, PA can be told how to interpret an ambiguous sequence (e.g. the two characters in a prenasalised stop are usually a single phoneme the language I'm analyzing).

Once the phonemes are being recognized properly, the histograms under the Consonant and Vowel charts will tell you how many times each phoneme occurs in your data. This is useful for helping you identify which phonemes to teach first in a literacy primer, and also which phonemes should probably be considered marginal phonemes. PA also has a powerful search/filter feature that helps you identify the various environments a phone can be found in (these searches can be stored), and PA can instantly generate Y charts and XY charts.

Probably the most useful feature PA provides is the "Minimal pairs" button. This allows you to find pairs of words that demonstrate contrast (either CIE or CAE) between any two sounds you've identified as phonemes. For example, you can look for pairs of words that prove the contrast between /d/ and /n/. You can even contrast /d/ with /nd/, assuming you've already told PA that prenasalized d is a single phoneme.

Phonology Assistant has few weaknesses and is very safe because it does not modify your data file(s). However, it does lack some features that phonologists might expect, since most of its functionality is on the phonetic side of phonology. For example, it does not provide a way to store or apply phonological rules. Still, if it did so it would have to assume a certain theoretical framework (generative rules would be simple but limited and outdated; an autosegmental or optimality approach would be more "current" but visually complicated.)

Likewise, it also is not able to handle morphophonemics very well, so if a certain phone (or phones) could come from either underlying phoneme(s) A or underlying phoneme(s) B, PA will treat them all as A, or all as B, unless you can use neighboring phones to consistently disambiguate them. Thus, if you represent /n/+/d/ and /nd/ in the same way phonetically, PA will rigidly interpret all occurrences in a single way (either all are two phonemes, or all are one phoneme). Still, handling morphophonemics well would require significant complexity--such as using a rule to refer back to a separate data field that is phonemic rather than phonetic, in order to disambiguate the phonetic field.

On the whole, Phonology Assistant gives the phonologist a convenient and solid toolset for sifting through phonetic data looking for patterns.