3.5 Sound-Change Rules

Listen for the condition

A sound-change rule links a sound change to an environment. The environment might be the next vowel, the previous consonant, the start of a word, the end of a word, or the stressed syllable. For DLAB-style aptitude work the important question is not just what changed; it is when the change happens. Capturing the condition is what separates a usable rule from a vague impression.

Compare, state, test

The core method has three steps:

• Compare several examples that share the pattern.
• State the rule in plain language, including the trigger.
• Test the rule on a new original item before committing.

For instance, in a practice system n might become m before p: san-pa becomes sam-pa, while san-ta stays san-ta. The rule is not "n always becomes m." The rule is "n changes before p." That precise condition is what protects you from distractors that contain the right sound in the wrong place.

What a complete condition contains

A condition statement should always include:

• The sound that changes.
• The environment that triggers the change.
• At least one no-change example showing where the rule does not apply.

No-change examples are not filler. They tell you where the condition fails to apply, and a rule that explains both the changed items and the stable ones is far stronger than one that only fits the changes.

Do not overgeneralize from one item

If you hear lom become lon a single time, you do not yet know whether m changes to n, whether all final sounds change, or whether that one word is simply irregular inside the drill. One example suggests; repeated examples confirm. Wait for a second instance before locking in a rule, because committing early to a wrong condition costs you every later item that depends on it.

State rules in small words

Use compact condition words: before, after, at start, at end, when stressed, when followed by. A rule such as "t becomes d between vowels" is far easier to apply under time pressure than a fuzzy memory that the word "sounded softer." Sound changes also interact with stress — an invented system might lengthen a vowel only when it is stressed. Ignore stress and the vowel pattern looks random; mark stress first and the rule becomes predictable. This is why audio prep must combine stress, rhythm, and discrimination rather than train them as isolated tricks.

Apply the condition, not the impression

When answer choices appear, do not let spelling override the heard pattern. The option that looks closest may break the rule. If the examples show n changes only before p, then a choice with m before t is probably wrong even if it feels similar. Review by writing the failed condition precisely: instead of "missed a sound change," write "changed n before t but the rule required change before p." That exact note retrains the ear to ask for the environment automatically, which is the real payoff of sound-change practice — and these remain original drills, never official DLAB items.

Common environments to scan for

When you suspect a sound change, run through a short menu of typical triggers rather than guessing. Most invented-language changes fall into a handful of categories:

• Assimilation — a sound shifts to match a neighbor (n becomes m before p, as the labels b/m/p share a lip closure).
• Final-position change — a sound at the end of a word weakens, drops, or hardens.
• Between-vowels softening — a hard consonant relaxes when surrounded by vowels (k to g, t to d).
• Stress-driven change — a vowel lengthens or a consonant strengthens only in the stressed syllable.

Knowing the menu means that when you hear a change, you can immediately test "is the neighbor doing it?" or "is it only at the end?" instead of cataloging every possibility from scratch.

A two-rule interaction example

The DLAB will sometimes stack rules so that two conditions operate at once. Suppose a system both lengthens a stressed vowel and drops a final vowel at the end of a phrase. The word ta-ku in isolation might surface as ta-KUU (stress lengthening), but at the end of a phrase as ta-K (final-vowel drop plus the lengthening no longer audible). A learner who tracks only one rule will hear chaos. A learner who marks stress first, then position can predict both surfaces. When two rules interact, resolve them in a fixed order every time so the output stays predictable rather than mysterious.

Building and self-checking a rule

After you state a candidate rule, run a quick falsification check: find an item where, by your rule, the sound should not change, and confirm it indeed stays. If even one example contradicts the condition, the rule is wrong or incomplete, and it is far cheaper to discover that during the compare step than after committing an answer. A rule that has survived a no-change check is one you can apply to new items with confidence.