User:Nicolas Perrault III/Reliable sources for Palaeolithic articles

The following is intended as a friendly note to Wikipedians that study the Lower and Middle Palaeolithic as a hobby and may not be aware of the nature of literature in this discipline. Secondary general-public sources such as newspaper articles, museum brochures and documentaries are as a soft rule considerably biased when discussing this discipline, especially with respect to the Neanderthal debate. They usually favour intelligent, creative, resourceful Neanderthals. This view is indeed advocated by a number of excellent researchers (e.g. Villa et al. 2014), and it is probably fair to say that it is gaining momentum. Yet it is still controversial.

Very few debates of the many that existed concerning Neanderthals 100 years ago have been solved. Two that have were of a phylogenetic nature and regarded whether Europeans descend mostly from Neanderthals (no) and whether modern humans interbred with Neanderthals (yes). We now do know that Neanderthals were habitual hunters of large game. We also know that they often used fire, but it is still debated whether they could produce it (yes: Shimelmitz 2014; no: Sandgathe et al. 2011). No other major behavioural debate has been resolved. This includes deliberate burial, symbolic behaviour, and language.

One often reads that scholars today know better than their early-20th century predecessors. For example, the Australian museum wrote in 2010 that "[t]he unfortunate stereotype of [Neanderthals] as dim-witted and brutish cavemen still lingers in popular ideology but detailed scientific research has revealed a more accurate picture". As much as one, myself included, may have a hard time believing humans so close to us with brains larger than ours were dim-witted brutes, it is simply not possible to reject this century-old stereotype on current evidence. The appropriate idea to share with the public is not that they were brutes nor that they could have designed Apollo 11. It is that Neanderthals could have been one or could have done the other, we just don't know for sure. A primary source that argues for Apollo in a well-reasoned extensively-researched manner is sharing a point of view, a normal part of the scientific process, not showing bias. But a journalist or encyclopedia that takes it for cash is always biased.

Here is a subjective appreciation in decreasing order of usability on Wikipedia (objectivity and verifiability) of sources in the study of Neanderthals in particular, and the Lower and Middle Palaeolithic in general:

Rooted sources
These sources support their statements with inline references.


 * Specialised encyclopedias for scholars: This is the quality and objectivity standards Wikipedia is aiming for. These encyclopedias generally express a neutral point of view and nearly every important statement is verifiable. This makes them massive collections of sources with little place for opinion (its readership has diverging view points). But entries are terse and their information much like in a dictionary may be fragmented into many small articles, which is not what Wikipedia is about. Most of their entries can be understood by anyone.
 * The Wiley-Blackwell Encyclopedia of Human Evolution (2013). My favourite. The only work I know on human evolution that I would call impartial. Unfortunately, it sometimes preserves its impartiality by not discussing a debate.


 * Popular science books with inline references:
 * In Search of the Neanderthals, by Chris Stringer and Clive Gamble (1993). A 31-year old book written with such skill and poise it remains unequaled as an introduction to the Neanderthals. You can get it for 3 pounds (5$) on abebooks.


 * Scientific articles: They are very well referenced but usually only present one side of the story. Citing a scientific article with no reference to the criticism it received or to papers suggesting an opposite view is the highway to bias. Its fast-lane is reserved for newspaper journalists.

Floating sources
These sources do not support their statements with inline references. You are invited to use them when rooted sources are unavailable, or as a supplement to them.


 * Specialised encyclopedias for students: The preface to the first edition of the Encyclopedia of Human Evolution and Prehistory (2000) sums up well the strengths and weaknesses of what's probably the best type of floating source: "Each entry has been prepared by a leading authority on its subject; and although every contributor was asked to represent all major points of view on the many topics that are the matter of dispute, each was left free to expound his or her preferred interpretation." These encyclopedias can hence be expected to show some limited bias. They are useful for what's uncontroversial, and for background research on the controversial. But the lack of inline references seriously hinders their usefulness as reliable sources to build an unbiased picture of anything subject to scholarly dispute, which includes almost everything related to behaviour.
 * Encyclopedia of Human Evolution and Prehistory, 2nd ed. (2004), edited by, Ian Tattersall et al. 54 authors.


 * Undergraduate textbooks: An undergraduate textbook is one of the good chances a generation of scholars has of influencing the next. It is where scholars will first try to correct the perceived wrongs of the previous generation. I recall wondering as an undergraduate whether the authors of our textbooks depicted Neanderthals in a more flattering light than evidence seemed to allow out of fear of being accused of perpetuating the century-old stereotype of the brute. These texts are similar in scope to encyclopedias for students, but are generally written, edited, and reviewed by far fewer authors, increasing the risk for bias.
 * How Humans Evolved, 7th ed. (2014), by Robert Boyd and Joan B. Silk.
 * World Prehistory and Archaeology, 3rd ed. (2013), by Michael Chazan


 * General encyclopedias: These are written with objectivity in mind, but it is hard to evaluate what information is up-to-date because they do not cite their sources inline. On occasion they present ideas as well-established when they are not, which is rarely the case in student textbooks. For example, Encyclopædia Britannica illustrates both these shortcomings at once when it states as if it had been shown beyond doubt that "about 40,000 years ago [...] Neanderthals in Europe began making a variety of more-advanced (Upper Paleolithic) tools from bone and stone that were frequently hafted". Not only has this never quite been the consensus (see e.g. White 2001), but the 40,000-year figure is outdated (see discussion) and the layers of the only site in which we have both this sophisticated material and Neanderthal skeletons was not-that-recently found to be heavily mixed with modern human layers (Higham et al. 2010, Mellars 2010). More recent research indeed suggests Neanderthals likely produced this material (Hublin et al. 2012, Ruebens et al. 2015, Roussel et al. 2016), but it is not yet certain (Ruebens et al. 2015). In no particular order:
 * Encyclopædia Britannica
 * Encyclopædia Universalis (French)
 * Brockhaus Enzyklopädie (German)


 * Articles in scientific magazines: Serious in intent and usually in content, articles in these magazines are often the vulgarisation of scientific articles. It's a good idea to cite both the scientific and vulgarised article at once, and treat the pair in your weighing of the facts as you would treat a scientific article.
 * Scientific American
 * New Scientist


 * Museum publications: These vary in quality, but they are often outdated, sometimes biased, and usually unreliable for details. For example, regarding skeletons discovered at the French site of La Ferrassie, on a page updated in 2016, the prestigious Smithsonian Institution writes that "[a] total of eight Neanderthal individuals -- including adults, children, infants, and two fetuses -- were found intentionally buried at La Ferrassie." This is wrong on two points, outdated on a third, and biased on a fourth. There is only one infant skeleton (La Ferrassie 4bis), not several. Those who found La Ferrassie 8 never claimed a burial. Following a 2002 article by Maureille, it seems highly likely that the remains called "La Ferrassie 4" are a mix-up of bones and actually part of Le Moustier 2, bringing the number of skeletons at La Ferrassie down to seven. Finally, intentional Neanderthal burials are controversial.


 * Popular science books with no inline references: These are similar in shortcomings to general encyclopedias, but are more likely to be subjective because written by fewer authors. The nature of the publication also means that the publishers put less pressure on the authors to provide a balanced point of view. The quality of these publications vary considerably.
 * Neanderthals Rediscovered 2nd ed. by Dimitra Papagianni and Michael A. Morse (2013).


 * Brochures and websites of universities: These are likely to reflect the school's research bias. Avoid, except for material of purely factual nature that you cannot get elsewhere.
 * Articles in general newspapers: Avoid whenever possible. In my experience of Palaeolithic Wikipedia articles, statements supported by general newspaper articles are more often biased or plain false than those appended with a tag. Of course, when supporting a claim with scientific articles or other rooted sources it's desirable as a courtesy to the layman to add a newspaper source making the same claim.
 * News or comments sections of scientific journals: These suffer from the biases of scientific articles (being written by the same people), but are not peer-reviewed and often give much fewer inline citations. The result is that they are often among the most lopsided and sensationalist sources available online. Here's an example I came across this afternoon:
 * "The Anthropocene Working Group of the [ ICS ] announced in August that over the next three years it will divide Earth's story into two parts: one in which humans are a geological superpower — an epoch called the Anthropocene — and the other encompassing all that came before our species had a major influence on Earth's functioning." (Nature News, 2016)


 * This is sensationalism. The geological time scale divides the geological history of the earth into eons (longest), eras, periods, epochs, and ages (shortest), with the Anthropocene suggested to be an epoch, within the Quaternary period (~2.6 million years), the Cenozoic era, and the Phanerozoic eon. No one is dividing the history of the earth in two.

Shortcuts to assessing bias
Understanding what a biased sentence looks like is worthwhile for two reasons. First, it allows you to avoid writing them yourself. Second, it allows making a rough initial guess of how trustworthy a source is. In Palaeolithic archaeology, the tone a writer uses (contained or overconfident) is a much better initial guess of impartiality than his or her authority. As we will see, some obscure popular science websites on occasion publish less biased articles than some scientific news published by highly respected scientific journals such as Nature. This does not mean you should cite obscure popular science websites or blogs, nor abstain from citing Nature News. It is simply an invitation to develop a sense for quickly assessing the extent to which an article is biased, or likely to be.

An intellectually dishonest author may try and pass a minority view as consensus. For example, in the lead of this Scientific American blog post, the author presents Neanderthal with an h as an alternative spelling to Neandertal, when it's actually the other way around. Neanderthal is used much more. The spelling one uses matters little, and this author may actually not be intellectually dishonest. But that it appears she's trying to fool you into thinking her chosen spelling is the most common one is a red flag to be twice as scrupulous when reading the rest of the article.

With three exceptions, it is probably a true generalisation that unrestrained, highly confident sentences on Neanderthal behaviour are non-negligibly biased. The exceptions:
 * Neanderthals made stone tools.
 * Neanderthals sometimes used [not made] fire.
 * Neanderthals were hunters.

Unless they had massive fur, much brown adipose fat, a considerably higher metabolism, or a combination thereof, they probably made some basic clothes as well.

Examples
Here are a few examples of how different news sources reacted to research on Neanderthal behaviour. Bias scores are of course subjective and depend on how well admitted a theory is in the scientific world (see next section). The article bias score weighs both the tone of how the research is presented and the weight given to criticism of the claim.

Neanderthal personal ornaments
A number of recent scientific articles have claimed Neanderthals used feathers and eagle talons for decorative purposes.


 * New Scientist: Did Neanderthals use feathers for fashion?


 * A legitimate question. But the tone of the title shows some limited bias in that it unfairly implies it was already known Neanderthals were close in cognition and sophistication to what would be required to use decorative feathers. A less biased title: Did Neanderthals have the sophistication to use decorative feathers? On the one hand, almost all of the article summarises the research conclusions of the scientific article to which it refers, giving it undue weight. On the other hand, to its credit, it does recognise the debate as to the abilities of Neanderthals.
 * Title bias score: 4/10
 * Article bias score: 3/10


 * Smithsonian: Do Feathers Reveal Neanderthal Brainpower?
 * Title bias score: 2/10
 * Article bias score: 6/10. Fair tone, but no review of criticism.


 * BBC: Neanderthals used feathers as 'personal ornaments'
 * The hefty bias of this title is only slightly toned down in the three last words of the article's first sentence: "Our evolutionary cousins the Neanderthals were harvesting feathers from birds in order to use them as personal ornaments, a study suggests."
 * Title bias score: 9.5/10, quotation marks save it from a 10/10.
 * Article bias score: 8/10, its tone is close to implying the finding is established fact, and there is no review of criticism.


 * Nature (News): Neanderthals wore eagle talons as jewellery
 * As mentioned above, the news section of scientific journals cannot be assumed to be reliable sources.
 * Title bias score: 10/10
 * Article bias score: 9/10

Historical example
Higher up we said that determining the bias is subjective and depends on scientific consensus. Here's how I would have scored the title's bias of this 1928 Scientific American article if it had been written at other times:

Neanderthal Man Not Our Ancestor

 * Title bias score in 1928: 1/10. Marcellin Boule had excluded the Neanderthal from Homo sapiens ancestry based on a now-discounted reading of the La Chapelle-aux-Saints 1 skeleton's anatomy, which he considered too brutish to be part of this line (because he thought the skeleton more primitive than Australian Aboriginals ). Today the data on which this early 20th century conclusion was based would be regarded as insufficient, but the title in its original publication year gets a low bias score because this conclusion was widely accepted. Unbiasedness is not about truth, but how well-accepted an idea is. The title's confident tone precludes it from a 0/10.
 * Unbiased title in 1928: Nearly all available evidence suggests Neanderthal man was not our ancestor.
 * Title bias score if this were written in 1988: 10/10. Such a claim presented as established fact in 1988 would have been seriously misleading. At this time, most researchers believed in the multiregional origin of modern humans, according to which Asian Homo erectus evolved into modern Asians, African Homo erectus evolved into modern Africans, and European Neanderthals evolved into modern Europeans. Some scholars, such as Chris Stringer supported the alternative theory of the recent African origin of modern humans.
 * Unbiased title in 1988: A minority of scholars believe Neanderthals were not the ancestors of modern Europeans.
 * Title bias score if this were written in 1998: 10/10. Beginning in the early 1990s, those claiming Neanderthals were the sole ancestors of modern Europeans were in the minority. Yet the minority was important enough for such a confident wording to score a 10/10 bias.
 * Unbiased title for 1998: Most scholars believe Neanderthals were not our ancestors.
 * Title bias score if this were written in 2008: 1/10. At this time there was a near-consensus for the Recent African origin of modern humans.
 * Unbiased title for 2008: Nearly all available evidence, including mitochondrial genetics, suggests Neanderthal man was not our ancestor.
 * Title bias score if this were written in 2024: 0.21/10. It is universally accepted since the early 2010s that non-African modern humans have a small number of Neanderthal genes (1.5-2.1%, Prüfer 2014).
 * Unbiased title for 2024: Genetic evidence strongly suggests very few of our ancestors were Neanderthals.

As of 2024, confident wordings on the subject of Neanderthals, especially behaviour, almost always reveal some form of bias.

What is knowledge?
Knowledge can be understood as belief that is both true and well-justified. Suppose I flipped a coin in front of a boy. It falls on heads, but I hide the coin with my hand so that he cannot see how it fell. The boy excitedly says: "I know it fell on heads! Every time I play heads or tails with my sister it falls on heads so it must be heads!" Would you say the boy "knows" the coin fell on heads? No you probably wouldn't. The boy holds a belief that is true and justified, but not well-justified. If I then showed the boy the coin and tried to explain him that he didn't really know it was heads, he might say "Whatever I was right!". But being right is not what matters, it's to be right and to have justified our belief appropriately.

In 2010, a genetics team led by Svante Pääbo at the Max Planck Institute of Evolutionary Anthropology published an important paper. In it they suggested admixture between Neanderthals and modern humans was the most parsimonious explanation for observed patterns in ancient DNA. Despite the great effort the laboratory expended over fifteen years to arrive at this result, they were careful not to make the claim unconditional. They conceived of an alternative, more contrived idea to explain their observations (genetic substructures from the time prior to the divergence of the Neanderthal and modern lineages). Because this alternative explanation — less parsimonious but still reasonable — had not yet been ruled out, hybridisation had not been shown nor proven in any meaning these words may have. Only strongly suggested. It was hence premature on behalf of news outlets to title articles


 * Neanderthals, Humans Interbred, DNA Proves (Seeker, 6 May 2010)
 * Neanderthal genome reveals interbreeding with humans (New Scientist, 6 May 2010)
 * Neandertal Genome Study Reveals That We Have a Little Caveman in Us (Scientific American, 6 May 2010)
 * Ancient DNA shows interbreeding between Homo sapiens and Neanderthal (Washington Post, 7 May 2010)

In 2012, the alternative explanation was reasonably ruled out. This does not forgive the bold titles of the above news outlets. They were right, but that's not what matters: in 2010, belief in hybridisation was not sufficiently well-justified to be said proven. The 2010 claim is now a member of this highly exclusive group of ideas we may consider knowledge. As discussed above, very few claims on Neanderthal behaviour have been shown to this degree of certitude. Hence the debate and hence the need for Wikipedians to discuss these claims with a greater amount of care than most news outlets do.

Suppose I tried a different trick with the boy's older sister. I take out a tricked coin both sides of which are heads. I flip it in front of her, hiding what side it fell on. The girl, her eyes glittering, then says: "I know it fell on tails! It was heads-up before you flipped it and it turned exactly three times in the air, so it must be tails!" The girl's belief is well-justified because indeed the coin turned three times in the air, but it is false and hence not knowledge. (I wouldn't call it very well-justified because the girl made the mistake of trusting me and didn't make sure the coin was fair.)

This definition of knowledge is the one closest to what we mean with this word in everyday life. But it is problematic in many parts of science because it requires access to the truth. In the strictest of senses, in spite of the enormous amounts of stone tools collected from their sites we do not know Neanderthals made stone tools, because we never saw them make them. Yet in practice anyone angrily reminding us that we do not know they made stone tools on this distinction alone would rightly be regarded as a pedant. We use the word knowledge in a closely approximate alternative sense to discuss things that are extremely well-justified — and hence highly likely to be true. Criminal justice lawyers call this proof "beyond reasonable doubt".

In all spheres of scientific research, accumulating the large amounts of evidence required to say that something is knowledge requires what Churchill might call enormous amounts of toil, tears, and sweat. Much more anyway than what is usually needed to make a guess at the most likely explanation.

Phrases to use with care

 * Reconstruction: Careless use of this word is a subtle way to introduce bias. It means to fairly accurately rebuild something as it was in the past. Artist interpretations are often presented as but are not reconstructions.
 * Advanced: Used in the absolute to qualify technology, this word is largely meaningless. How advanced is it? It is preferable, for instance, instead of writing that culture X had "advanced clothing", to write that it had highly insulated, quilted, sewn, and tailored clothing that was usually woven, but also knitted, and crocheted. Used in a relative way, such as "X had more advanced weapons than Y", "advanced" does carry meaning. But in these cases, as when used in the absolute, it is preferable to show than to tell: "X used stone-tipped wooden spears, while Y only threw rocks".
 * Complex: As "advanced", usually meaningless, and when meaningful, there are usually better alternatives. Of course, it is legitimate to use "complex" in contexts where it is given a special meaning. For instance, in a scientific article, a "simple tool" can be defined as one made of one part (a stick) and a "complex tool" defined as one made of two or more parts (stick, glue, and stone).