Questions, evidence, models and their limitations

When you can no longer see the solid

Suppose a recorded demonstration shows sugar crystals being added to water, followed by a clear solution. “The crystals are no longer visible” describes an observation. “The sugar was destroyed” proposes an explanation. “The sugar molecules became dispersed through water” proposes a different explanation. The observation alone does not choose between them.

Turn the uncertainty into a question: what evidence would distinguish loss of visible crystals from loss or transformation of sugar? We can ask whether the dissolved material can be identified, whether an appropriate separation recovers it, and whether the mass measurements agree with the boundary of the system. This lesson analyses descriptions and records; it does not require a home experiment.

Make a prediction that could matter

A useful explanation has consequences beyond retelling the observation. If sugar molecules remain in solution, an appropriate method should detect sugar there. If it has been converted into other substances, analysis should instead find evidence for those products. The method must actually distinguish the alternatives; measuring colour would be a poor test when all relevant samples are colourless.

Choose conditions and comparison samples that make the result interpretable. A water-only comparison can help show whether a measured signal is associated with adding sugar. Records should state sample identity, amounts, conditions and the method’s limitations. A test that cannot detect an expected amount cannot support a strong claim that the substance is absent.

State what is being measured

A system boundary identifies the matter included in your accounting. If water evaporates from an open cup, weighing only the cup and its remaining contents excludes the water that entered the surrounding air. A lower reading does not show that matter ceased to exist.

For ordinary chemical processes, mass balance treats total mass as conserved within a closed material system to the accuracy of typical chemical measurements. Count incoming and outgoing matter when the system is open. This practical statement does not say that every measured reading is exact, or that mass and energy can never be related. Later physics explains the finer limits.

Read agreement at the instrument’s resolution

Consider an invented teaching record: the same closed container and contents give readings of 150.00 g before a change and 150.01 g after it. Its displayed increment is 0.01 g. The difference is 0.01 g, but the increment is not a complete uncertainty estimate. Calibration, repeatability and other effects also need consideration.

A careful conclusion is that these two readings are close; deciding whether their difference is meaningful requires the measurement uncertainty. Do not replace them with identical numbers to make the expected law look tidy. Agreement is evaluated using the actual observations and the capabilities of the method.

A model earns its usefulness through testing

A model is a selected representation of a system, such as the particle sketch of dissolving sugar. It leaves some details out so that a particular relationship is easier to reason about. State the assumptions, identify its predictions and compare them with relevant evidence. A successful prediction supports the model for that purpose; it does not verify every omitted detail.

If the evidence disagrees, check both the model and how the evidence was obtained. A recording error, a different sample, an uncontrolled condition or a missing interaction might matter. Repeated agreement across suitable independent methods can make an explanation stronger. There is no automatic ladder on which a guess turns into a law just by surviving a fixed number of tests.

Worked examples

Try it yourself

1. Which is the observation: “the crystals are no longer visible” or “the sugar molecules were destroyed”?

   Show the answer and reasoning

   The crystals are no longer visible.

   The second statement proposes a cause and needs evidence that can distinguish it from other explanations.

2. An instrument cannot detect the small concentration expected in a sample. Does a zero signal establish that the substance is absent?

   Show the answer and reasoning

   No.

   Absence of a detectable signal must be interpreted with the method’s detection capability and relevant interferences.

3. Subtract the two invented readings 150.00 g and 150.01 g. Is this difference alone enough to establish a real mass change?

   Show the answer and reasoning

   The difference is 0.01 g; the readings alone are not enough.

   You also need measurement uncertainty and conditions. The displayed increment is only one part of the instrument’s behaviour.

4. A particle model correctly predicts one result. Must every detail drawn in the picture be physically exact?

   Show the answer and reasoning

   No.

   Success supports the prediction within its tested scope. Drawing conventions and omitted interactions still need appropriate interpretation.

What to carry forward

• Keep observations separate from explanations.
• Seek evidence that can distinguish the relevant alternatives.
• Account for matter crossing the boundary of a measured system.
• Measurement resolution and measurement uncertainty are different.
• Use models for stated purposes and revise them when evidence requires it.