Sound change is regular, systematic, and exceptionless across all words meeting a given phonological environment — the Neogrammarian hypothesis. Grimm's Law describes the systematic consonant shift distinguishing Germanic languages from other Indo-European branches (e.g., Latin /p/ → Germanic /f/, as in 'pater' vs 'father'). The comparative method identifies cognates (words sharing common ancestry) across related languages and uses regular sound correspondences to reconstruct earlier proto-forms. Apparent exceptions to regular sound change typically reflect analogy, borrowing from a related language, or a conditioning environment not yet identified.
Work through Grimm's Law correspondences step by step across cognate sets. Apply the comparative method to a small constructed data set of three or four related languages. Practice distinguishing true cognates from loanwords using sound correspondence tests.
From articulatory phonetics, you know how speech sounds are produced — where in the mouth, with what degree of voicing, how air flows. From historical linguistics, you have the broader picture: languages descend from common ancestors and change over time through regular processes. Sound change is where those two threads come together. The central claim of the Neogrammarian school is striking: sound change is exceptionless. Every word that meets a given phonological environment undergoes the change, without exception. This isn't just a strong claim — it's a methodological commitment that makes historical reconstruction possible.
Grimm's Law is the canonical demonstration. In the 19th century, Jacob Grimm documented a systematic pattern: where Latin, Greek, and Sanskrit have certain stop consonants, the Germanic languages (English, German, Dutch, Gothic, etc.) consistently show different ones. Latin *pater* / English *father*, Latin *tres* / English *three*, Latin *canis* / Greek *kyon* / English *hound*. The shift follows a pattern: voiceless stops became fricatives (/p/ → /f/, /t/ → /θ/, /k/ → /h/), voiced stops became voiceless (/b/ → /p/, /d/ → /t/, /g/ → /k/), and voiced aspirates became voiced fricatives or stops. What makes this a law rather than a curiosity is that it applies systematically across the entire Germanic vocabulary — not just these examples, but every eligible word.
The comparative method turns this regularity into a reconstruction tool. If you have several related languages and you find that Latin shows /p/ wherever English shows /f/, wherever Sanskrit shows /p/, wherever Greek shows /p/, you can infer a Proto-Indo-European */p/* that remained stable in most branches but shifted in Germanic. Cognates — words sharing common ancestry — are identified by testing whether the differences between them follow the established sound correspondences. English *night*, German *Nacht*, Latin *nox/noctis*, Greek *nyx/nyktos*, Sanskrit *nakta* — these look different on the surface, but their vowels and consonants correspond exactly to what the sound change rules predict.
Apparent exceptions are where the analysis gets interesting. When a word seems to violate a regular sound change, the Neogrammarian move is not to abandon the rule but to ask what process is masking it. Analogy (paradigm leveling, where irregular forms get regularized on the model of related forms), borrowing (a word from a related language that underwent a different history), or an unrecognized conditioning environment (the sound change was more narrowly specified than initially thought) account for most apparent exceptions. This methodological discipline — treating regularity as the default and explaining exceptions rather than allowing them — is what gives historical phonology its scientific rigor and its power to reconstruct languages spoken thousands of years before writing.