Handle ambiguous and conflicting mineral evidence

Three outcomes need three different interpretations

A mineral key can leave several candidates, one candidate or none. Several candidates mean the current evidence does not discriminate enough within the list. One candidate means only that it is the sole listed match under the assumptions. No candidates means the observations and this key cannot all be reconciled as stated.

These outcomes are not percentages of certainty. Two remaining names do not imply a 50–50 chance. The browser has no prior frequencies, error rates or likelihood model. It performs exact comparisons of supplied categories.

Preserve an ambiguous result

Choose Specimen A and include the cleavage pattern but not the angles. The specimen is scratched by a hardness-4 reference and has three dominant cleavage directions. Calcite and halite both match. “Unresolved between calcite and halite in this key” is a more accurate statement than picking a name without further evidence.

The missing angle observation addresses the remaining distinction. A right-angle result supports halite within the list; a non-right-angle result supports calcite. Before revealing it, write both conditional predictions. This makes the test plan explicit and prevents interpreting every possible outcome as support for the preferred answer.

A missing observation differs from a negative observation. “Angles not recorded” is not “angles are not right angles.” Likewise, failure to inspect a broken surface is not an observation that cleavage is absent. Keep these distinctions in a real notebook.

Inspect a conflicting record

Select Specimen E. At the visual stage, all four references are compatible. At the scratch stage, “not scratched” leaves quartz. Adding the record of three dominant cleavage directions excludes quartz, leaving no candidates. Including non-right angles afterward does not repair the conflict with the earlier scratch result.

This record was deliberately constructed to expose incompatible evidence. It is not a newly discovered mineral and does not establish which observation is wrong. The pattern could motivate several hypotheses in real work: a surface mark was misread, different grains in a mixed specimen were tested, a crystal face was mistaken for cleavage, or the true mineral is outside the candidate list.

The workbench retains all included observations. It does not choose the label matching the greatest number of properties or silently discard the inconvenient result. An additional column cannot resurrect an excluded candidate unless you explicitly revise the underlying record or key, which these controls do not do.

Design a targeted recheck

For Specimen E, separate three possibilities before planning more work. First, if the scratch observation is unreliable, inspect the same labeled grain using an appropriate reference procedure and distinguish a groove from a surface residue. Second, if the cleavage interpretation is unreliable, inspect repeated parallel breakage surfaces and distinguish them from growth faces. Third, if multiple grains were involved, map the tested locations and determine whether the observations describe the same material.

These are proposals for supervised follow-up, not procedures executed by the browser. Additional microscopy or compositional analysis may be appropriate when the restricted key is inadequate. A proposed test should state which outcomes would favor each explanation and which uncertainty would remain.

A repeat can improve observation confidence while adding little discriminatory information. If both calcite and halite predict the same scratch category, repeating that category may check reliability but still cannot distinguish those two candidates. Test validity and test usefulness answer different questions.

Keep the scope of the key visible

The reference list includes only four minerals and idealized representative observations. An actual natural specimen may be an aggregate, altered, coated, poorly crystalline or a mineral absent from the list. A single surviving candidate therefore depends on both trustworthy observations and an adequate candidate set.

Consider an unfamiliar mineral with the same four recorded categories as a listed candidate. This key would be unable to distinguish them. That is a limitation of the evidence, even if the software filters its table perfectly. You would need another informative property or a more complete analysis.

The reference disclosure shows the profiles used in filtering. It is intentionally public learning material. A successful answer after reading it is guided practice, not a hidden identification test or evidence that you can operate laboratory equipment independently.

Transfer: repair a contradiction without hiding it

Suppose a new record says “not scratched by the hardness-4 reference” and “three dominant cleavage directions at right angles.” The key returns no matches. A student chooses halite because it agrees with the cleavage observations and deletes the scratch result. Explain the error and state a better next step.

Check the mineral-evidence transfer

Halite conflicts with the supplied scratch record, while quartz conflicts with the supplied cleavage record. Choosing halite and deleting evidence hides that conflict. Preserve both observations, check whether the same grain and a valid scratch comparison were used, and investigate whether the key omits the true mineral. If a repeat justifies revising a record, retain the original, correction and reason rather than pretending the discrepancy never occurred.

Capture one ambiguous case and one conflicting case. In the reflection, identify the smallest set of observations causing each result, propose a discriminating check and avoid assigning a probability that the evidence does not support.

Reference properties are linked in the four-mineral key and in the mineral-properties lesson. The conflicting specimen, reasoning comparisons and transfer exercise are original; they are not recorded measurements.