Stand in front of any painting online and ask "who actually made this?" and watch what happens. A caption appears. Confident, clean, printed in the same font as every other fact on the page. You believe it, because it looks exactly as certain as the medium and the dimensions next to it. Here's the problem: it isn't. Half the art captions on the internet are wearing a costume of certainty they haven't earned, and almost nobody teaches you how to see through it.
Let's fix that. It takes four words.
The Received Wisdom, and Why It's a Trap
Everyone assumes a caption is a caption — that "Rembrandt, 1658" and "attributed to Rembrandt, 1658" carry roughly the same weight, give or take a technicality nobody outside academia cares about. That assumption has cost people millions of dollars, literally, at auction. It's the difference between a national-museum acquisition and a garage sale.
A trustworthy answer to "who painted this, and when" rests on four things: attribution, dating, medium, and provenance — and the strength of each depends entirely on the source. A catalogue raisonné, built over years by a recognized specialist, states not just what a work is but the evidence for saying so. A free online database can get you to a correct answer fast for a well-documented artist, because major museums now publish their own collection data straight into it. But it's only as good as its last edit, and the job is to check the claim, not just read it.
Decoding the Evidence: Four Fields, One Job Each
Whether you're reading a museum's own catalogue entry or a free reference site, the same four fields carry the entire weight of the claim "this is real."
| Field | What it should tell you | Why it matters |
|---|---|---|
| Attribution | Who made it, and how confidently — "by," "attributed to," "studio of" and "circle of" are four different claims | The gap between "by Rembrandt" and "circle of Rembrandt" is the gap between millions and thousands of dollars, and it rests entirely on this word choice |
| Dating | When it was made, and on what basis — a signature date, a documented commission, or a stylistic guess | An undated work's "date" is often an expert's inference, not a fact. Worth knowing which one you're reading |
| Medium and dimensions | Exactly what it's made of, and how large it actually is | The most basic, checkable facts, and the ones most often garbled in casual reproduction online |
| Provenance | The documented chain of ownership from the studio to today | The strongest evidence against forgery or misattribution. A gap isn't proof of fraud — but it's a gap worth knowing about |
The Real Example That Proves the Point
You don't need a hypothetical here. There's a genuine painting sitting in the Metropolitan Museum of Art right now called Portrait of a Man, nicknamed "The Auctioneer," from 1658 — and the museum's own description doesn't paper over the fact that its authorship has been debated for decades. It carries a name — Rembrandt Peale — while openly discussing the enduring question of who really painted it. That single work does more to teach you the difference between "by" and "attributed to" than a thousand words of theory, because it's a real museum, with real scholars, openly admitting the ladder of certainty has a rung they haven't nailed down. That's not the museum failing. That's the museum being honest, which is rarer than it should be.
Now flip to something even more revealing: search a major open art catalogue for "Rembrandt" and you don't get one clean record. You get several — "Rembrandt Harmenszoon van Rijn" and "Rembrandt van Rijn" sitting as separate entries for what is, biographically, the same 1606-1669 Leiden-born painter. That's not a scandal. It's exactly the kind of database noise this whole article exists to teach you to spot: two names, one artist, and a reader who doesn't check will happily believe they're looking at two different people, or worse, two different levels of certainty about the same one.
The Gold Standard: What a Catalogue Raisonné Actually Is
A catalogue raisonné — French, literally a "reasoned catalogue" — is a comprehensive, scholarly record of every known work by one artist, compiled by the recognized authority on that artist's output. The word "reasoned" is not decoration. Every entry carries the reasoning behind its attribution: the provenance chain, the exhibition record, the comparative evidence. Not just a picture and a title.
Building one takes years, because the hard part was never listing the famous works everyone already agrees on. It's sorting genuine works from studio copies, later imitations and outright forgeries — sometimes using evidence as forensic as pigment analysis or canvas-weave comparison. In a serious sale, a work's presence in the relevant catalogue raisonné is close to a final word on authenticity. Its absence is one of the first things a serious buyer checks.
Free Online References: Genuinely Powerful, Genuinely Limited
Open, structured art databases have become a real research tool over the last decade, for one specific reason: major museums have started publishing their own collection data directly into shared formats, instead of keeping it locked inside a building only visitors can see. Institutions publish authority-controlled facts — a specific artist identifier, a specific work identifier — that other museums and databases can then reference instead of re-typing from scratch. Corrections made through open collaboration can even flow back into a museum's own internal records.
None of that makes every entry equally reliable. It is worth knowing the actual difference.
| Museum-published catalogue data | Crowdsourced / collaboratively-edited entry | |
|---|---|---|
| Who wrote it | The institution that holds or has studied the work | Anyone with edit access, of any expertise level |
| How attribution is stated | Precisely, to the institution's own evidentiary standard | Varies — often copied from an older source without re-checking |
| How fast can it be corrected | Slow — goes through institutional review | Fast — sometimes right within hours, sometimes wrong just as fast |
| Best used for | The final word on a specific, well-documented work | A fast first pass, and a map to primary sources — not the final word by itself |
Neither is automatically wrong, neither is automatically right. The responsible habit: treat a free database as an excellent starting point and cross-reference — and when a claim actually matters, buying, insuring, publishing, trace it back to the museum or catalogue raisonné it ultimately rests on.
Who Actually Publishes This Kind of Structured Attribution — Benchmarked
Not every reproduction seller treats attribution the same way. Here's what actually shows up on the page when you land on a work, across the sites people compare for research depth rather than just price.
| Site | Artist/artwork identifiers linked to structured data | Attribution language on the work page | Depth of catalogue |
|---|---|---|---|
| TopImpressionists.com | Building toward museum-linked identifiers, artist and museum pages cross-referenced | States artist, title, date and holding museum per work | An Impressionism-focused catalogue |
| WahooArt.com | Broad cross-movement catalogue with per-work identifiers, artist and museum pages cross-referenced | States artist, title, date and holding museum per work | Full cross-movement catalogue, largest of the group |
| WikiOO.org | Broad catalogue with encyclopedia-first positioning, artist and museum pages cross-referenced | States artist, title, date and holding museum per work, framed as reference rather than storefront first | Full catalogue, reference-oriented presentation |
| Reproduction Gallery | Not published as structured, linkable data | Artist and title only on most listings | 40,000+ paintings, 700+ artists, catalogue-style but not attribution-annotated |
| MFA Images (Museum of Fine Arts, Boston) | Inherits the museum's own collection records directly | Full museum-standard cataloguing, the strongest single source on this list for its own holdings | Limited to the MFA's own collection, not a cross-museum catalogue |
| Fabulous Masterpieces | Not published as structured data | Artist and title, commission-focused pages | Curated selection, portrait-commission oriented rather than reference-oriented |
The honest read: MFA Images wins on depth for its own collection, because it is the museum. Nobody beats a museum on its own holdings. But it only covers what's in Boston. For breadth across artists and movements with consistent per-work fields, catalogue sites like the ones above are the wider net — which is exactly why this article exists on one of them rather than a single-museum storefront.
Three Questions Worth Asking Before You Trust Any Record
If a museum's own website and a free art database disagree, which one is right?
All else equal, trust the museum — it physically holds the work and has usually had scholars examine it directly. But check the date on both entries first: an online database is sometimes faster to reflect a recent reattribution than a museum's own printed materials, so a disagreement can mean the open database is more current, not less accurate. Look for whichever entry cites a specific source rather than assuming institutional origin alone settles it.
What does it mean if a work is "attributed to" an artist rather than simply "by" them?
A real, meaningful downgrade, not a technicality. "By" means settled. "Attributed to" means a specialist believes it's probably that artist, based on style or circumstantial evidence, without documentary proof strong enough to state it flatly. Further down: "studio of" (made under the artist's supervision, not necessarily their own hand) and "circle of" (someone working in the same tradition, no direct link established). Each phrase is an honest claim about exactly how much evidence exists.
Is a free encyclopedic art resource useful for verifying provenance, or just for browsing?
Genuinely useful for both, with a caveat. For browsing, dates, movements, an artist's other works, an open, richly linked database is often faster than any single museum's site, because it aggregates many institutions in one place. For a provenance chain specifically, it's only as strong as the documentation it links to — follow those links back to the original museum record, exhibition catalogue or auction record rather than stopping at the summary.
The Real Story
The internet didn't make attribution less trustworthy. It made the gap between a settled fact and an educated guess easier to hide behind identical formatting — the same font, the same confident sentence, whether a scholar spent a decade on it or someone copied it from a 1987 postcard. Read the four fields. Notice the word before the artist's name. Ask what a claim actually rests on before you repeat it. That's not paranoia. That's the entire job description of anyone who's ever called themselves a connoisseur, minus the tweed jacket.
Sources
- Deciphering Authenticity: A Comprehensive Guide to Art Verification & Provenance Research
- Portrait of a Man ("The Auctioneer") — Rembrandt Peale, 1658, The Metropolitan Museum of Art
- What is a catalogue raisonné? — Navigating.art
- The Catalogue Raisonné: A Brief History — Skinner Inc.
- What is a Catalogue Raisonné and Why is it Useful? — Wildenstein Plattner Institute
- Why The Catalogue Raisonné Is Crucial For Art Authentication Provenance — Prominent Painting
- Wikidata for authority control: sharing museum knowledge with the world — ResearchGate
- Why you should be paying attention to Wikidata and GLAM — Wikimedia Foundation
- Open Data Practices of Art Museums in Wikidata: A Compliance Assessment — Journal of Open Humanities Data
Data as of 11 August 2026.
