3.4 Phonetic Discrimination

Make small sounds count

Phonetic discrimination is the ability to hear the difference between sounds that feel close at first contact. On the DLAB the goal is not to name sounds with technical symbols; it is to notice whether a small difference is consistent enough to control the answer. A final t instead of a final d may matter. A short vowel instead of a long vowel may matter. A missing sound may matter. Treat each tiny difference as candidate evidence until the examples confirm or rule it out.

Start at the edges

Final consonants are the easiest sounds to lose under time pressure, so train them first. A practice set might use mip, mit, and mik. If words ending in p name tools and words ending in t name actions, the final sound is the rule, and the shared mi is background. Beginnings matter too: lum versus slum turns on the initial s, and noka versus nok turns on a final vowel. Invented-language rules often live at the edges, exactly where rushed listeners stop paying attention.

Voicing pairs

Voicing is whether the vocal cords vibrate during a consonant. Pairs like p/b, t/d, and k/g share a mouth position but differ in voicing, which makes them feel deceptively similar. In a drill, pa and ba must be treated as different if the examples make that difference meaningful. You never need to say the word voicing on test day — you only need to hear that the pair is not identical.

Use broad, stable vowel labels

Vowels are hard because learners force unfamiliar sounds into English categories. Use broad labels that only need to be stable enough to compare: high vs low, front vs back, tense vs relaxed, or simply vowel A versus vowel B. If the same vowel appears in every example of a category, it becomes a candidate rule. Consistency, not precision, is what you are after.

Minimal pairs into patterns

A minimal pair changes exactly one feature: ta versus da, ko versus ku, nal versus nas. Once you can hear the pair, embed it in a larger pattern — every item ending in s could be plural while every item ending in l is singular. These remain original practice-style examples, not official DLAB content.

Two-column review

Review misses with a two-column note. In one column write not heard when you could not perceive the contrast; in the other write not used when you heard it but did not apply it. Not heard errors need slower discrimination drills. Not used errors need better rule comparison. This split prevents vague review and makes each session efficient. The larger skill is disciplined attention: the DLAB deliberately uses unfamiliar, gibberish-like material to measure language-learning potential, and the calmer you compare small differences, the less you guess from familiarity.

Long versus short vowels and consonant length

Many invented systems exploit length because English barely uses it to distinguish words. A held vowel (maa) can mean something different from a short one (ma); a doubled consonant (at-ta) can differ from a single one (a-ta). English ears tend to ignore these because length is not contrastive in English. Train it explicitly: practice ma / maa, su / suu, and ato / atto until you can flag the held sound reliably. If the examples show that long vowels mark one category and short vowels mark another, length becomes the deciding feature, and a choice that copies every segment but uses the wrong duration is wrong.

Position-sensitive listening

The same sound can behave differently depending on where it sits. A k at the start of a word may be easy to hear, but a k between two vowels can soften and become slippery. Train your ear to check sounds at the start, in the middle, and at the end as separate listening targets rather than assuming a sound is identical wherever it appears. This habit pays off directly in the next section, where conditional sound changes depend on exactly these positions.

A discrimination warm-up routine

Before an integrated session, run a sixty-second warm-up of pure minimal pairs with no rules attached: pa/ba, ta/da, ka/ga, mi/ni, ko/ku, lo/lu, mip/mit, nok/noka. Answer only same or different and check yourself. This isolates raw perception so that when rules enter, a miss can be cleanly blamed on rule use rather than on a tired or untuned ear. Warming up the discrimination system first is the cheapest way to keep your not heard errors near zero during the harder integrated drills.

Think of it the way a musician runs scales before a piece: the goal is not the warm-up itself but a tuned instrument that no longer fails on the basics when the real demands arrive.