We live in world
where AI is playing an ever increasing, potentially problematic role in our
lives. It may be no surprise to many tool collectors, but several recent
experiences have convinced me that this technology has the potential to significantly
enrich our understanding of the historical characteristics and origins of many
of our antique tools. This is especially true for “what’s it’s”, but even for
tools that we feel we know quite a bit about.
I recently used
Google’s Gemini AI tool to investigate the identity of a dozen of my “what’s
its”. This involved posting a photo and letting Gemini have a go at it. I
believe ultimately it was able to identify most, but not all of them. The
program will sometimes offer several reasonable options and leave it to you to choose
one. Often Gemini’s first suggestion is clearly wrong. The next step is key.
You must come back at Gemini with a reason its description is likely incorrect
and add additional characteristics of the tool in question which the program
appears not to have paid attention to. Often the program will ask you specific
questions regarding some aspects of the tool. This back and forth may continue
for several rounds until an identification is provided that makes sense. At the
end of each of Gemini’s paragraphs it usually provides a link to a source it
has apparently utilized, or which may provide pertinent information.
Another way that
AI can be of use to antique tool collectors is its ability to expand one’s understanding
of a known tool. The following is
an example using a large pair of early European wing dividers in my collection.
Gemini was able
to focus on a range of the tool’s characteristics. This discussion went on for
quite a while and generated 11 pages of dialog when printed out. The dialog began with a question I raised
regarding the ways to distinguish between dividers from the 1600’s and those of
the 1700’s. Gemini came up with five differences, summarized the comparison and
provided photos and sources. I then brought in the example above to see what
Gemini would say.
Above is the ball joint with a
flared base characteristic of a 17th century tool.
Simple end to the
wing as opposed to the 18th and 19th century decorated
end along with a “lorgnette” thumbscrew (round, flat head with pierced hole).
The hole allowed the user to insert an iron rod (“tommy bar”) to firmly tighten
the screw.
The unthreaded “through
hole” on the backside of the thumb screw boss (housing) found on some early
dividers is apparently a purposeful detail. It provided a way for shavings from
early taps or later debris to escape as well as a way to remove a snapped
thumbscrew.
The program
suggested an origin of these dividers from the Netherlands, Germany or England.
When I questioned why France was left off the list it replied that I had caught
a major omission and that, based on several of the detailed characteristics, the
dividers were likely French and falling into the Louis XIV era from 1643 –
1715.
We then shifted
to the fixed point of attachment of the wing. Instead of being riveted in place
Gemini stated it was secured by a “hot weld”, a more complex 17th
century technique than the later use of a rivet.
The top and
bottom of the legs for three inches above and below the wing bosses are
decorated with ogee/cyma swirls. This decoration and its location is apparently
characteristic of the High Baroque period between 1660 and 1690. This further
solidifies the dating of the tool.
Decorations are
hot forged not cold file work. This is associated with the late 17th
century.
The 21” length is
typical of dividers used in timber framing, shipbuilding and millwrighting.
The program even explained
the occurrence of an apparent forging accident that resulted in a “red-short
split delamination” and the correction with a fire-weld repair.
After summarizing
the tools characteristics and providing numerous references, Gemini offers to
provide comparative examples.
All of Gemini’s
observations are drawn from existing information in on-line sources so, given
enough time and motivation, you could probably generate a similar analysis
without the use of AI. Google’s AI program simplifies this task significantly. You
must maintain a critical eye on Gemini’s analysis since the complex algorithms rely
on the validity of data from the sites it draws from and incorrect assumptions
are certainly possible. In a later query regarding when wing dividers first
appear Gemini claimed it was in the late 1500’s. When I brought up the painting
entitled “God the Geometer” from the 1200’s the program claimed the wing on the
pair of dividers God is holding was not a wing, but part of the orb being
measured. I replied that the painting clearly showed a wing passing through the
leg of the dividers and some form of adjustment knob on the other end. Gemini admitted
that these were in fact wing dividers and thanked me for the correction.
If used
critically AI offers a powerful research option that can be used in many ways
to increase our knowledge of the tools in our collections.
Note: AI would
probably have been more than happy to help write this blog or write it itself
(which it didn’t). Also, beware as the
program tends to flatter the person asking the questions and will even admit to
using flattery when questioned!
