Artifacts and craft is the study of objects as evidence: what a thing was made of, how it was shaped, who handled it, and what the wear on it says about use. The core method is reading the object against the record. A musket stock, a coin, a uniform button, a pottery sherd, a canteen — each carries a material signature and a manufacturing signature, and the two rarely lie in the same direction. The craft question ("how was this made?") and the artifact question ("what does this object prove?") are the same question asked from opposite ends.
Reading an object: material, method, and wear
Every artifact yields three separate lines of evidence, and they must be gathered in that order: material first, then method, then wear. Material tells you what the maker could do. Method tells you what the maker chose to do. Wear tells you what happened after the maker let go. Skipping to wear — the romantic part, the "this was carried at Gettysburg" part — is how provenance stories get built on objects that were never there.
Material identification is the most objective step and the one most often rushed. A "brass" button that is actually gilt over a base metal, a "leather" strap that is a later replacement, a "hand-forged" iron fitting with visible die flash from a drop hammer: each is a different object with a different date range. The tests are cheap and mostly non-destructive — a magnet, a jeweler's loupe, a close look at fracture surfaces, sometimes a scratch test in an inconspicuous spot. What goes wrong is that a collector buys the story first and the object second, then tests only to confirm.
Method follows from material. A sand-cast brass buckle has a slightly rough, matte back and a crisp front; a die-struck one has sharp edges on both faces and often a faint seam. A hand-thrown vessel shows throwing rings on the interior; a slip-cast one shows a mold seam. These are not decorative details — they are the manufacturing fingerprint, and they date an object more reliably than any paper tag attached to it decades later.
Wear is where interpretation begins, and where it most often overreaches. Genuine use wear is asymmetric and cumulative: a trigger guard thinned on one side, a coin flattened at the high points of the relief, a canteen strap worn through at exactly one point. Reenactment wear and deliberate "aging" tend to be symmetric, evenly distributed, or concentrated where a hand would never actually rest. Learn to look at the high points and the contact points, and the difference usually shows.
The cost of doing this properly is time, not money. A loupe, a magnet, good light, and a notebook will carry you through most of it. The cost of skipping it is that you build a collection of stories rather than a collection of objects, and stories do not survive scrutiny.
Provenance: what a chain of ownership can and cannot prove
Provenance proves continuity of ownership, not authenticity of manufacture, and the two are constantly confused. A documented chain that runs from a soldier's family to a museum to a dealer to you establishes that this object passed through those hands. It does not establish that the object was made when the chain says, or that the soldier ever carried it. Those are separate claims requiring separate evidence.
The strongest provenance is contemporaneous and independent: a period photograph showing the object, a wartime letter describing it, an inventory list with a physical description specific enough to exclude similar items. The weakest is a typed card from a dealer, a family oral tradition with no corroboration, or a "papers" file whose earliest document is a twentieth-century appraisal. Oral tradition is not worthless — it often points in the right direction — but it degrades with each retelling, and the degradation is invisible in the retelling itself.
Watch for provenance that arrives with the object rather than before it. If the documentation surfaced at the same moment as the sale, treat the chain as unproven until each link is checked independently. The classic failure mode is a single unattributed letter that names the object in general terms, folded into a file that then gets cited as "documented." One document does not make a chain; a chain is at least three links from independent sources.
The honest trade-off: rigorous provenance is expensive and slow, and it will cost you objects. Plenty of genuine pieces have no paper trail at all, because the families that kept them did not keep records. A paperless object with strong material and wear evidence is often a better buy than a papered object with weak material evidence — but only if you can live with the uncertainty, and only if you price it accordingly.
Dating by typology, and where typology fails
Typology dates an object by comparing it to a known sequence of forms, and it is the workhorse of artifact study because it needs no laboratory. The logic is simple: manufacturers change designs over time, sometimes for reasons of cost, sometimes for reasons of fashion, sometimes because a supplier changed. If you can place a form within that sequence, you can bracket its date.
The method is strongest where production was centralized and documented — military accoutrements, standardized coinage, marked ceramics, patent-dated hardware. It is weakest where production was local and informal, which is most of what ordinary people actually owned. A locally made tool can persist in the same form for generations, and a "late" form can appear early if a local smith happened to prefer it.
The standard failure is treating a typology as a date rather than a range. A form that "appears in 1863" did not stop being made in 1863; it often continued for decades, and the later examples are usually cruder. When you see a form that matches a wartime pattern but the finish is rough and the proportions slightly off, the likeliest reading is a later copy, not a rare variant. Typology gives you a window, not a timestamp, and the window is usually wider than the seller wants it to be.
Cross-dating helps: when typology, material, and a maker's mark all agree, the bracket tightens considerably. When they disagree, believe the material and the mark over the form. Forms get copied; alloy composition and die work are much harder to fake convincingly at scale.
Conservation and the ethics of cleaning
Do not clean an artifact until you know what the cleaning removes, and assume that in most cases the answer is "evidence." Patina, corrosion layers, residues, and staining are data. A bronze coin's green crust may preserve the original surface beneath it; strip the crust and you strip the surface with it. An iron object's rust layer often holds the shape of the original surface after the metal itself has gone.
Mechanical cleaning — a scalpel, a brush, a wooden pick under magnification — is reversible in the sense that it stops when you stop, but it is not reversible in what it takes off. Chemical cleaning is worse: it acts everywhere at once, it keeps working after you think you have rinsed it, and it can mobilize salts that will continue to attack the object for years. The conservator's rule is minimum intervention, and the reason is not aesthetic fussiness — it is that every intervention destroys information that cannot be recovered.
Stabilization is different from cleaning, and it is usually the right choice. Removing active corrosion, controlling humidity, isolating the object from acidic paper and from off-gassing plastics, and storing it in stable conditions will do more for an object's survival than any amount of scrubbing. The costs here are real: archival storage materials, silica gel, a hygrometer, and a stable room are not free, and a proper treatment by a conservator can exceed the object's market value. For a common piece, good storage is the rational choice; for a rare one, professional treatment is worth it.
What goes wrong when people clean anyway: they brighten a surface that should be dull, they remove the tool marks that proved hand manufacture, they destroy the residue that would have identified the object's last use, and they convert a documented artifact into an anonymous decorative object. The object survives; the evidence does not.
Fakes, reproductions, and the honest gray zone
Most inauthentic artifacts are not clever forgeries. They are reproductions made for reenactors, souvenirs, or film props that later entered the market without their labels, and they are detected by exactly the material and method tests described above — wrong alloy, wrong manufacturing process, wrong wear pattern. The elaborate forgery is rare; the mislabeled reproduction is everywhere.
The tell is usually process. Genuine period metalwork was made with tools that leave specific marks: hammer facets, file chatter, inconsistent thickness, hand-cut threads. Modern reproduction tends to be too uniform, too crisp, and too consistent in thickness, because it was made by machines that cannot help being consistent. A "hand-forged" item with no hammer marks, or with marks applied after the fact in a regular pattern, is a reproduction regardless of what the tag says.
The gray zone is real and worth naming honestly. Some reproduction is legitimate and openly sold as such; some is period-correct in every visible respect and is sold without any claim either way. The problem is not the reproduction — it is the missing label. When you buy, ask what the seller's basis for a date claim is, and be suspicious of any answer that rests on "it came from an old collection." Old collections contain reproductions too, sometimes older than the collector.
For the buyer, the practical defense is to price uncertainty explicitly. An object with no independent dating evidence should cost what an undated object costs, not what a documented one costs. The gap between those two prices is the seller's story premium, and paying it is a choice — sometimes a reasonable one, if you want the object for what it is rather than for what it proves.
Putting an object in historical context
An artifact earns its place in history when it can be tied to a dated event or process, and the tying is where most interpretation goes wrong. The dates themselves are firmer than people assume. The U.S. Civil War ran from 1861 to 1865, and the U.S. National Park Service figure for Civil War dead stands at 750,000. The western Roman Empire is commonly dated as ending in 476 CE, when Romulus Augustulus was deposed. World War II ran from 1939 to 1945, with the Pacific War continuing until Japan's surrender on September 2, 1945.
Those dates matter to artifact study because they set the brackets within which a material or manufacturing signature becomes evidence of a particular conflict or period. A cartridge box that matches a wartime pattern is Civil War-era material only if the leather, the stitching, and the hardware all fall inside the bracket; if the hardware is a later alloy, the box may be a postwar assembly using wartime parts, which is a different and much more common object than people expect.
| Context | Date bracket | What the bracket lets you claim |
|---|---|---|
| U.S. Civil War | 1861–1865 | Material and hardware consistent with wartime production; 750,000 dead (National Park Service) |
| Western Roman Empire | Ends 476 CE | Coinage, pottery, and metalwork dated to the western sequence before the deposition of Romulus Augustulus |
| World War II, Pacific | 1939–1945; Pacific War to September 2, 1945 | Industrial production signatures and issue markings tied to a specific theater and end date |
The caution is that a date bracket is not a provenance. Knowing an object was made within a window does not place it at any particular event inside that window, and the gap between "period" and "present at" is where most inflated claims live. A Civil War-era canteen is period material; a canteen carried at a specific battle requires a chain of evidence that most objects simply do not have. The honest statement is the narrower one, and it is usually interesting enough on its own.
The reward for doing this carefully is that the object becomes a genuine source rather than an illustration. A button with the right alloy, the right die work, and wear in the right places tells you something about how a soldier's kit was actually made and used that no general history will. That is the whole point of the craft: the object, read properly, is evidence — and evidence is worth more than atmosphere.





