8.4 Eliminate Options With Rule Evidence

Elimination is still reasoning

A multiple-choice format tempts candidates to hunt for the answer that looks right. A stronger approach is to eliminate the choices the evidence forbids. This keeps the DLAB skills in play — comparison, rule extraction, memory, and application — instead of pattern-matching on appearance. Because the DLAB does not subtract beyond a missed point, narrowing four options to two and guessing is always better than freezing on a red item.

Start elimination with a short list of what the examples prove. If the examples show subject-object-verb order, drop choices that use subject-verb-object unless another marker justifies the change. If the examples show a negative suffix, drop affirmative choices for a negative target.

Practice-style example, not official DLAB content:

Target: The pilot does not carry the radio.

The correct construction needs pilot, radio, carries, and negation, in the observed order. Evaluate the choices against one rule at a time:

One rule at a time saves memory

Instead of holding four complete answer strings in working memory, run them through a fixed checklist: order → required marker → actor → object. Each pass discards options cheaply. This is especially valuable late in the 126-item session, when holding four full strings at once is exactly what overloads a tired mind.

Elimination also rescues partial knowledge. Suppose you cannot tell whether a suffix marks tense or negation, but you know the target must share a feature shown in a given example. You can still remove every choice that omits that feature, even before you fully solve the system.

Traps to avoid

• Style-based elimination. Longer is not better, shorter is not cleaner, and a word that resembles English is not safer. Repetition means nothing unless the examples make it mean something.
• Overfitting. If two rules both fit the examples, do not eliminate a choice that violates only your favorite unsupported theory. The overfitting trap — locking a rule onto too little evidence — quietly removes correct answers. Eliminate only with evidence the examples establish.
• Over-elimination to zero. If your checklist leaves no options, you applied a rule the examples did not prove. Reopen the most aggressive cut.

Train elimination directly

For some drills, do not try to solve fully at first. Ask only, "Which two choices can I remove, and exactly which rule does each violate?" Then finish. This builds the speed the DLAB demands without abandoning logic. During review, inspect your eliminated choices: if you removed the correct answer, name the false rule that did it; if you left an obviously wrong answer standing, name the evidence you missed.

This discipline matters because the DLAB feeds selection and placement decisions under service and agency policy, and the published category cutoffs (I=95, II=100, III=105, IV=110) mean a handful of avoidable misses can change which language door opens. Evidence-based elimination is how you keep careful reasoning alive when the clock and the 126-item count are pressing on you.

A repeatable elimination checklist

Turn elimination into a fixed sequence you can run on any constructed-language multiple-choice item, in order, stopping as soon as one option remains:

1. Order check. Does the option follow the word order the examples prove? Drop any that reorder without a justifying marker.
2. Required-feature check. Does the target demand a specific marker (negation, plural, tense)? Drop options that omit it.
3. Role check. Are the actor and object the ones the target names, not a swapped pair? Drop role reversals.
4. Extra-piece check. Does the option add a syllable or affix the examples never justify? Drop unsupported additions.

Run the checks in this order because order and required-feature violations are the cheapest and most reliable to spot, so they cull the field fastest while your working memory is freshest.

The partial-knowledge advantage

Elimination shines when you cannot fully solve the system. Imagine you are unsure whether -su marks plural or past tense, but the target clearly requires whatever -su does (it parallels a given example). You may not know the rule's name, yet you can still delete every option lacking -su and keep every option that has it. That single move can take a four-option item down to two — a coin flip you have biased heavily in your favor — without ever resolving the ambiguity.

That last row is the discipline that separates strong test-takers from overfitters: when the evidence genuinely underdetermines the rule, you stop eliminating and guess among the survivors rather than inventing a tiebreaker. On the DLAB, a defensible 50/50 guess beats a confident cut built on a rule the examples never proved — and it costs far less time, which you will need for the items still ahead.