Two Portraits, Two Labels, and a Student Who Wants to Look It Up
Two paintings are being prepared for a school exhibition. Label A reads "Italian, 15th century, tempera and gold on poplar panel"; Label B reads "Northern European, 16th century, oil on canvas." Students observe that A shows a fine network of vertical cracks running with the grain, a greenish tone under the flesh areas, and a halo of tooled gold. B shows visible canvas weave, deep transparent shadows in the drapery, and distinct brushmarks in a sleeve. One student asks, "Why are we guessing? Can't we just look up who painted it?" A second student points out that Label B could have been painted in Italy, since Italians used oil too.
1. For each label, identify the material evidence that supports it and the evidence that should make students cautious about treating the label as settled fact.
Read the worked answer
Label A (tempera and gold on poplar, Italian 15th c.):
- Supporting: a poplar panel splits and cracks along the grain, producing the fine vertical network described; egg tempera is applied in small hatched strokes and builds slowly, favoring the crisp, linear handling typical of the period; a green underlayer beneath flesh is consistent with a layered tempera method; tooled gold leaf with punched or incised decoration is a period-typical technique and physically sits above the paint film.
- Caution: cracks along the grain also occur on any panel painting, including Northern ones. Gold leaf was used widely across Europe. A later copy or a workshop production can imitate all of these. Panel paintings are also frequently transferred to a new support, which can erase or alter the crack pattern and the surface.
Label B (oil on canvas, Northern European 16th c.):
- Supporting: canvas weave is visible, indicating a flexible support; oil allows translucent glazing, which produces the deep shadows described; visible brushmarks in a sleeve fit oil's working properties better than tempera's.
- Caution: the second student is right - oil on canvas was used in Italy by the 16th century, so the medium does not discriminate between regions. A canvas may have been lined onto a new canvas, so the visible weave may not be original. Dark glazes are the most frequently abraded and restored layer, so "deep transparent shadows" may be partly a restorer's reconstruction.
The general principle to teach: material evidence narrows a range of possibilities; it rarely names a hand. Each label is a hypothesis with a confidence level, not a fact.
2. How would you answer the student who asks, "Can't we just look it up?" - in a way that teaches historical method rather than dismissing the question?
Read the worked answer
Take the question seriously; it is the right instinct applied to the wrong authority.
Concede the useful part: yes, we consult catalogues, museum records, and scholarship - no historian works alone. Then show what those sources actually contain:
- Attributions are provisional and get revised. Works move between "anonymous," "circle of," "workshop of," and a named artist as evidence changes. A student who looks it up may find two sources that disagree.
- Provenance has gaps. Ownership records can be incomplete, forged, or lost to war and fire. A gap is not proof of anything.
- Labels carry institutional incentives - a named artist sells tickets and loans; a museum may be slow to downgrade an attribution.
- Copies, workshop production, and forgeries exist, which is exactly why material evidence matters: pigment analysis, infrared reflectography revealing underdrawing, and dendrochronology (tree-ring dating of the panel) can distinguish a period piece from a later imitation.
Then hand the method back: "If you look it up, you now have a claim. Your job is to tell me what evidence would have to be true for that claim to hold, and what evidence would break it." That converts look-up from an endpoint into a source to be evaluated - which is the standard being assessed.
3. Design a short in-class task that forces students to reason from the objects rather than the labels, and describe what a strong response would look like.
Read the worked answer
Task: "Label removal." Cover both labels. Give pairs a worksheet with three columns - Observation (what I can see), Inference (what it suggests), Confidence (high / medium / low, and why). Prompt: "Write the most defensible label you can for each painting, and list the one piece of evidence that would most change your mind."
A strong response would:
- Separate observation from inference explicitly ("vertical cracks parallel to the long edge" → "a rigid support, likely a panel that has moved with humidity" → "medium confidence, since transfers can mimic this").
- Use medium properties correctly - hatching and layered opacity pointing toward tempera; translucent glazing and visible brushwork pointing toward oil - without claiming the medium proves the region.
- Assign different confidence levels to the two works, and justify the difference.
- Name a specific disconfirming test: pigment analysis for a date range, infrared imaging for underdrawing, tree-ring dating of the panel, or thread-count and weave analysis for the canvas.
- Address the strongest counterargument, e.g., that the Italian oil hypothesis is live for Label B.
Weak responses restate the original label, treat one piece of evidence as decisive, or use "it looks old" as the reasoning. Score the observation/inference separation and the willingness to lower confidence - those are the skills the standard names.