Laser authentication

[pkfrey]

I know the authentication business is not the best topic, there are many varied opinions, some positive, but mostly negative. I am just trying to keep current on what's going on. Does anyone know if the laser treatment to authenticate artifacts is being accepted and is considered reliable? It doesn't matter who or what business performs these tests, I am just trying to find out the current status of this treatment among collectors. Thanks for any opinions.

[painshill]

Hi Paul

There was a bit of previous discussion on the forum here:

http://forums.arrowheads.com/forum/g...authentication

“Laser” is a broad term in this context, encompassing laser-induced breakdown spectroscopy (LIBS), laser-induced fluorescence (LIF), infra-red laser spectroscopy (ILS) and Raman spectroscopy among others.

None of these techniques can “authenticate” an artefact. They essentially provide a chemical and/or structural fingerprint at atomic or molecular level in the form of a spectral map with the absorbance or reflectance at particular wavelengths. That kind of data can then be compared to the spectral map for “known” lithics or other raw materials and also empirically correlated to the potential age of an artefact made from that material. In the wrong hands, it provides unreliable information for which unwarranted fast-and-loose claims are currently being made by some vendors.

What we have here is a classic problem relating to the distinctions between data, knowledge and learning.

If I gave you the numbers: 37, 42, 50, 65 and 70 then that would just be “data.” If I then told you that these were the measured heights in inches of five skeletons found together in a single burial context and that they were all female apart from the largest skeleton, then it becomes “knowledge.” You then might use that knowledge in combination with other knowledge you have accumulated from elsewhere to generate a hypothesis that would result in “learning.” For example, you might conclude (based on knowledge of human growth patterns) that the three smaller females were 3, 5 and 8 years old at time of death and that they might represent the offspring of the two larger (adult) individuals. Perhaps a family all died at the same time in a tragic accident, or were stricken by an epidemic of something infectious. You might further conclude (based on typical social patterns of partnership) that the female adult was most probably a little younger than the male adult; that they partnered up in their late twenties or early thirties; and waited a year or so before starting a family. That would enable you to estimate the ages at death of the adults too.

All is well and good (on average), providing you reliably know where the skeletons were found and that the location represents a likelihood of decent nutrition and conventional social norms. Without that certainty, your comparative data would need to embrace the possibilities for all ethnic groups, poor nutrition and other possible deviations. Also, if it so happened that (unbeknown to you) one of the small skeletons represented an adult midget, and one of the large skeletons was an adopted refugee child from a tall African ethnic group, then your “learning” would fall apart.

In a simplistic sense, that’s how these kinds of laser-based techniques are used. They don’t measure “age” any more than a wooden rule measures the age of an individual at time of death based on examination of his skeleton as per my example above. They generate data which, when expressed as knowledge, allows you to make an estimate based on other reliable knowledge that you already have.

I spent about 5 years of my career working with instruments that measure near infra-red reflectance, laser-induced emission spectra and such. Techniques have improved dramatically since those times and you can now get sophisticated instruments which combine several principles within a single piece of equipment. However, the scientific improvements largely relate to the measurement accuracy of the data, rather than the accuracy of the prediction from the data.

Even then, it’s important that data is gathered under standardised conditions by laboratories working to accredited standards and using the same methodology. Otherwise it’s not comparable. For example, much of the seawater temperature data relating to “global warming/climate change” is fundamentally flawed arising from what is known as the “bucket correction” (the British Navy historically used a wooden bucket for sampling; the more recent measurements principally from the US Navy are from metal buckets; and most modern measurements come from water flowing through ships’ turbines). Similarly, the US historical air-temperature data is largely from measurement stations in open areas away from human disturbance… which often meant somewhere in the corner of a weather station’s parking lot with a large expanse of tarmac. It’s not directly comparable to more recent professional data, or that from other countries, without some kind of theoretical correction.

The principles for prediction from “laser techniques” have not really changed. You still need a calibration curve, it still needs to include a robust data set from “known” samples and it still relies on statistical techniques such as linear backwards stepwise regression (which used to take forever before we had Pentium-chip computers with massive RAM) to make a prediction from the “unknown” data. In some cases, the non-professionals in this field aren’t taking the comparisons that far. They’re just overlaying an unknown spectral map on a known spectral map and saying: “look, the peaks match”. The real improvements for prediction come from the completeness of the known data set and proper statistical analysis. The longer these techniques have been in existence, the better that data set becomes because there has been more “donkey-work” to gather and correlate known data. But it’s a massive hurdle for accuracy.

For example, I was at one time using near infra-red reflectance (NIR) for the analysis of fish samples. I was only interested in “cod” and had put several hundred known data sets into my calibration, using samples from all oceans where the fish is commonly found. However, I also had to include reference samples from fish pre- and post- spawning, fish held on ice for various time periods post-mortem, and samples that been previously frozen. That’s what increases the robustness of the prediction and – even then – it became more robust when I extended the data set to include other gadoid fish such as haddock and whiting, which might turn up as unknowns and mistakenly presumed to be cod.

Similarly, in archaeological applications, the accuracy of any prediction critically depends on the robustness of the calibration set. Obviously, an artefact found in a swamp in Louisiana will not generate the same data set as an artefact of the same age found in the Arizona desert, even if they were both made from the same lithic. The comparative data set(s) would need to include all possible petrological variations for the lithic and all possible burial and exposure conditions for the artefact, as well as including heat-treated lithics and examples with faked patina from chemical treatments.

I would say that, even as the techniques and reference databases continue to improve, this kind of methodology is only likely to be reliable for artefacts made from readily identifiable lithic sources which have been recovered from undisturbed depositions in regions with a known geochemical and climate history. Other than that, the techniques can only provide data in support of that indicated by style/type, stratigraphy and cultural context (ie, is the age prediction consistent with the general hypothesis?)... or for exposing outright fakery.

You might be interested in this article from Dr Gramly, which he seems to have rather mischievously dated 14^th April 2019:

http://www.insightlae.com/drgramlyaso.pdf

[pkfrey]

Rodger, Thank you very much for responding to this. I appreciate your insight , opinion, and the references. I did read the publication by Dr. Gramly. I've known Mike for quite a while, and although I don't always agree with his assessments, he is extremely knowledgeable in the field of archaeology. BUT, what I had to do was look up some of those long intelligent words you used! Where did you get all that education? I'm impressed, I thought I knew a lot, not everything, but quite a bit about archaeology. I don't even know how to spell some of these words! My impression on measuring the levels of radon from the interior of an artifact would be inconclusive, because, the rock the artifact was made from, which I call the " parent " stone, would already be, geologically, 100s of thousands of years old, so of course there will be high levels of radon in the artifact made from the parent stone. So actually, it's not conclusive that this test proves authenticity in any artifact. There are to many variables that may interfere with accurate results. 2019? I wonder what Mike is waiting for!! Anyway, I got a lot of information from your response, and I truly respect your opinion, and taking the time to respond. Thanks, Paul

[painshill]

Thanks Paul... and sorry about the big words, but this area of science is a lot more than just measuring radon levels!

Essentially I'm a biochemist and worked as such for a number of years, but I also spent a spent a large chunk of my early career running the applied physics section of the research centre for a multi-national corporation. Although none of the testing I conducted was of an "archaeological" nature, many of the techniques I was using were essentially the same as those used for forensic archaeology. I was doing or co-ordinating animal species testing on blood and protein residues (as is employed to determine what a blade artefact might have been used for); DNA testing to check for genetic modifications (as employed to determine ethnic origins); optically stimulated luminescence testing on silica residues (as used to determine exposure ages for lithic artifacts); gas-liquid chromatography and mass spectrometry (used to identify trace amounts and residues of unknown organic compounds); instrumental colour measurement (as used for lithic matching); deformation, impact and tensile testing of materials, including man-made glasses (as employed in lithic analysis/experimental knapping) and a whole bunch of other stuff, including the near infra-red work I referred to.