The Rift Valley is beautiful and amazing, well beyond my powers of description.  Driving north from Nairobi on the old road is spectacular.  We are now, thanks to some advise from Meave Leakey, well above the present shoreline of Lake Elmenteita near a quarry.  The lake level used to reach up past here, in the good old days, hundreds of meters higher than the present levels of Lake Nakuru and Lake Elmenteita.

     And this ancient shoreline is what made Kariandusi into another version of what we saw at Olorgesailie yesterday:  here, too, there is a Leakey-excavated sea of handaxes.  Except that here, you can easily see one layer atop another.  And then another, like a layer cake.  My cousin, primed by seeing Olorgesailie and now knowing what to look for, is gratifyingly amazed.  Kariandusi is about the same age as Olorgesailie, with estimates ranging from 700,000 to a million years old, right in the middle of Homo erectus time.  Together, Olorgesailie and Kariandusi give you a wonderful lesson in why erectus toolmaking is so enigmatic.  And why climate changes moving lakeshores might be so central to understanding Homo erectus.

     As at Olorgesailie, there is a fine one-room museum at Kariandusi, another branch of the National Museum of Kenya which runs this site.  The museum here has the same hominid skull casts seen in Nairobi (and in museums around the world).  It also has a nice short version of how the Rift Valley formed in geological time, the floor dropping down between two cracks (the two escarpments we see mark, naturally, the two big north-south fault lines) as the rift widened.  The walls of the valley have a lot of fossils exposed in some places.

     The museum appears to train the guides well.  If some of what they say about handaxes is implausible, it’s only because they are faithfully repeating what archaeologists have said over the decades (sometimes, one suspects, with tongue in cheek).

The Kariandusi museum has a poster full of the proposed uses of the handaxe.  It shows you the extremes to which archaeologists have been pushed, trying to find some definitive use for the strangely-shaped handaxe which is the defining hallmark of Homo erectus culture.  They all have uses for it – and they’re probably mostly correct, too – but none adequately explain why its shape is what it is, or why that shape escapes the usual cultural drift for close to 1.8 million years.

     Many modern archaeologists are suspicious that they might be dealing with a blind-men-and-the-elephant situation when it comes to the handaxe, that they possess only part of the truth and are making erroneous extrapolations.  They just don’t know what to replace or supplement them with.  Because of the history of inadequate explanations, most have become totally skeptical of any new explanation that is offered for the handaxe, even by fellow archaeologists.  Been there, done that.

     The handaxe poster starts with “For butchering.”  Yes, one edge of a handaxe could be used for clearing meat from the bones.  But why is the back edge sharpened too, the one the hand has to wrap around?  I’ve never heard an answer to that one.  And why use an elaborate tool when a simpler one usually suffices?  Why use such a symmetric, hard-to-make tool for a purpose that one of Glynn’s half-rocks or random flakes would suffice for?  It’s overkill, as the phase goes.  The handaxe sure wasn’t designed for just defleshing.

     “For digging up edible roots” reads the next panel.  That’s even more difficult when the implement is sharpened all around the perimeter, because you tend to pound on roots.  You’d break many handaxes that way (not to mention cutting the palm of your hand).  And, in any event, overkill again.  There are some sturdy bifaces called “picks” that might stand up to cutting roots loose, but here I am concerned with the classic, enigmatic handaxe shape, not what broken or tumbled versions of it might secondarily prove useful for.

     “For scraping animal hides.”  (Clothing, so early?  Well, just assume they mean skinning the beast to get at the limb meat.)  I’d sure prefer something with a comfortable grip, myself.  And one of Glynn’s flakes or half-rocks would work pretty well for skinning.  It doesn’t even have to be particularly sharp for skinning, so long as you maintain tension by lifting the hide loose as you separate skin from fascia.  When I learned my surgery, one of the big surprises the first week was how handy dull tools were, that you often wanted to use the dull side of the scissors to spread, or the blunt handle of the scalpel to scrape, rather than the sharpened surfaces.  Only beginners think that the scalpel is the quintessential surgeon’s instrument.  Once you get through the skin somehow, surgery is mostly about spreading and separating along natural planes, such as between fascia and skin, where a sharp edge is a royal nuisance most of the time.

     “As a core from which tools were struck.”  That’s a good possibility, at least for the larger handaxes that lack fine retouching and show cavities where a flake used to reside.  But it doesn’t explain the bilateral attempt at symmetry, or the tendency toward a point.

     “As an anvil.”  Now there’s a trivial one.  Any flattish rock that was carried around would do, of course.  No design needed.  But if you were carrying around a designed rock that was flattish and sturdy enough – sure, why not use it as an anvil, should you need to soften up a root or crack a nut?

     “Thrown discus-style as a weapon.”  Now we’re talking.  Throwing has great possibilities in human evolution.  But the museum’s artist got it all wrong, as the cartoon shows an over arm throw at a lone antelope, the handaxe’s pseudo-point conveniently facing toward the animal, miraculously stabilized in that unlikely orientation without the aid of a spear shaft.  But at least “discus-style” was included in the caption, what I think is the key to the puzzle.

So you can see why it’s been called the “Swiss Army Knife of the Paleolithic.”  If no single use can justify its design, maybe it is a composite tool!  Or so the story goes (that phrase “jack of all trades and master of none” comes to mind).  Yet, ironically, the name “handaxe” is a misnomer because using it as a handheld axe is the least likely of its proposed uses.  In its classic form, it is sharpened all around its perimeter.  It is thus likely to bite any hand that held it, if striking something with any force.  So put scare quotes around “handaxe” from here on, whenever I forget them.

     Furthermore, almost none of the proposed uses account for its distinctive shape.  Certainly not simple butchering.  It could easily have been used as a core, from which smaller flakes were made, and there is some evidence for this use.  Unfortunately, many handaxes have so much fine finishing – tiny chips removed here and there – that it is clear that such handaxes were something more than merely a core for thumb-sized flakes.  Yes, hacking, scraping, and cutting are possible uses (and there is even some edge-wear supporting woodworking) – but the handaxe wouldn’t have done those jobs better than Glynn’s broken rocks with a fortuitous edge.  And they (and the Oldowan tools) are sure a lot easier to make, and they all continued to be made (the handaxe didn’t replace them).

     And why is the handaxe flattened and bilaterally symmetric, with a suggestion of a point? (It is, however, seldom sharpened to a penetrating point, the way you see in the spear tips and arrowheads in the last ice age.)  None of the Swiss Army Knife explanations ever get around to addressing such handaxe issues.

     Even worse, this classic shape persists for nearly 1.8 million years, even into Africa’s Middle Stone Age about 100,000 years ago.  It is seen in Africa, Europe, and now even in China.  It stays about the same shape,  regardless of what local rock is used.  That really suggests we’re missing something important.  To conserve the classic form through the vicissitudes of cultural drift and local fashion, there must be something that keeps the shape drifting back toward some functional optimum.

     But for what function?  It’s surely not the Swiss Army Knife functional collection, as combination devices are surely subject to drifting styles, with features being omitted when they are no longer supported by the culture.

     I suspect that I wouldn’t have told you all this if I didn’t myself have a candidate for what the prime use was, the one that none of the Swiss Army Knife uses adequately addresses.  It builds on some experiments that Eileen O’Brien did back in 1979, the most unsettling undergraduate thesis I have ever seen.  And since it meshes nicely with the waterhole aspect of the savanna story that I’m trying to get across, perhaps I should tell the updated version of my handaxe story.

When I first hatched my handaxe scheme back in the 1980s, someone immediately attached the name “Killer Frisbee” to it.  Just try throwing a handaxe, even with the initially horizontal plane of rotation favored by most modern Frisbee throwers, and it will soon turn on edge and come down vertically.  Happens even to experts, not just me (I repeated O’Brien’s experiments a decade later, also employing a well-trained discus thrower).  Doesn’t matter how you throw it – overhand, sidearm, underarm – the same thing happens.  “Handaxes” are vertical-plane spinners.

     Unlike round Frisbees, a handaxe will not lead you on a merry chase after landing, rolling along forever.  The handaxe’s “point” will shortly bury itself in the ground and bring the implement to a shuddering halt.  Is this good for something?  Some context must be missing, and it’s not likely to be a prehistoric game of paleofrisbee.

     The context, I suggested back at that Portugal meeting, is when throwing such an object into a tightly-packed herd visiting a waterhole, especially during a drought when their choices in watering sites are very restricted and they try the safety-in-numbers strategy.  Then the puzzle parts seem to fall together.

     I don’t think for a moment that the handaxe was a general-purpose hunting projectile (the archaeologists are rightly skeptical of that use).  I think that you can understand the handaxe only in the special-purpose context of closely-packed herds at the waterhole.

Let me start with a few words about shoreline predation, surely a topic of general importance for hominid evolution, not merely the handaxe enigma.  The big cats, as well as various other carnivores, have all discovered that lakes, rivers, and waterholes are a good place to hang out.  That’s because the grazing animals eventually have to visit a water source.  Wait for the animals to come to you.  Saves chasing them, and all that effort.  Hominids would have realized the same thing.  (Recall what the bank robber said when asked about why he robbed banks:  “Because that’s where the money is.”)

     Of course, after a few successes, the herd gets a little wary and picks another watering site.  But sometimes there isn’t another choice.  Some waterholes are obligatory, as all the others within walking distance have become mud holes, suitable only for elephant beauty baths.  And so each evening, the herd, packed tightly together for mutual protection, comes down to the waterhole.  The tails of the waiting predators begin to twitch in anticipation.  Sometimes there were tailless predators instead, of the hominid variety.

     Being on the outer rim of the herd is not the best place to be.  Packing tightly does reduce the percentage of the herd that is exposed to predation.  The bigger the herd, the bigger the perimeter – but the perimeter grows more slowly than the area, and so, in the larger herds, only a few percent are actually exposed to an attack.  With only a dozen animals, half of the herd is exposed.

     An upright hominid might have been able to get closer to the herd than the familiar four-legged carnivores.  But, as I earlier mentioned, the two-legged advantage doesn’t last long.  The herd will increase its approach distance – the distance from which they start moving away from you, if not running away.  And besides, what’s the hominid going to do, anyway – get close enough to club a hapless animal?

     So there’s the scene:  tightly-packed herd visiting waterhole, two-legged wannabe predator standing at some distance, without even a tail to twitch.  What happened next?

The entry-level solution, I said back at the Portugal meeting, was to throw a tree branch into the midst of the herd (chimpanzees like to wave and fling branches, so maybe our ancestors did too).

     The herd will panic, of course, wheeling around to flee.  The branch lands in their midst (towards the rear is best), perhaps hitting an animal, perhaps not.  But soon an animal trips over it or gets entangled, knocked off its feet by adjacent stampeding animals.  It tries to get back on its feet but another knocks it down in the confusion.  Pretty soon it gets halfway back on its feet, but it is too late.  Two-legged predators arrive and grab it.  Maybe they club it, maybe four just cooperatively hold a leg each in the chimpanzee manner while a fifth chews on its throat.  One way or another, they make a meal of it.

     The bigger and more tightly packed the herd, the better the technique works.  The usual surface-to-volume ratio advantage of herds and schooling is turned on its head by this simple invention, of lobbing a branch over the top and into their midst.  The bigger the herd, the easier it is to disable one in the middle.

     Pretty soon, herds grow wary – but they form bigger and more tightly packed groups in response, making the technique work even better.  The local trees are denuded of convenient branches.  While waving and flinging branches are clearly favorite scare tactics of chimpanzees (they have not been seen to use accurate throwing for predation per se), they will resort to clods of dirt and rocks when branches are in short supply.  Rocks, the hominids might have noticed, will go farther than anything else.  The thrower can launch while still outside the approach distance.  But what does the rock buy them?  Few animals will trip over it when fleeing.  With a beginner’s throw, knockouts are unlikely.

     Ah, but the herd is, remember, tightly packed in this scenario.  Even with the beginner’s side-of-the-barn throw, a lob can’t miss hitting some animal somewhere.  If the rock is big enough and the animal small enough, even a rock landing on the flank will bowl over the animal.  Were it not for the rest of the herd running past at close quarters, the downed animal could likely get up and run away in time.  Few would be injured enough to stay put.

     But the herd is running past, and attempts to get back up in time may fail.  Another hominid meal is procured.

     Rocks are more reusable than branches, which get trampled into mere sticks.  Are some rocks better than others for this tightly-packed-herd use?

Suppose you had an outcrop of sedimentary rocks that was layered, so some rocks were pre-flattened in the manner of slate.  Throw one of those and you’d get some of the frisbeelike aerodynamics.  Goes farther.  Lands on edge, too.  Sometimes on an animal’s back.

     That has an interesting neurological consequence.  All mammals have flexion reflexes, used to withdraw the body part from further harm.  If you step on a thorn, you quickly withdraw your foot and shift your weight to your other foot, all before your brain even takes notice of the thorn.  Most archaeologists know this, but what they might not know is that withdrawal reflexes also work for damage to other parts of the body – such as the back or flank.  If a grazing animal felt something sharp on its back (as from overhanging thorn trees, with acacias common hereabouts), it would tend to sit down quickly.  The big cats may have discovered this reflex tendency to crouch, given how they leap onto an animal’s back and dig their claws in. 

     So flat stones with sharp edges might get a reputation for being more effective.  If you are also knocking off sharp edges from such a stone to butcher the animal, it might have some sharp edges for next time.  Two uses for the price of one.

     The preferred projectile is now basically flat with sharp edges, somewhat symmetric for less tumbling en route.  But it doesn’t necessarily have a point.  Why add a suggestion of a point, without making it really pointed for penetration?

     Some would have been semi-pointed, of course, perhaps as a consequence of all the flaking used for butchering.  Or perhaps just from breaking off a segment (some would be trampled by the herd, after all, and then recovered).  And the hominid hunters might have noticed that a bit of a point made the stone more effective in bringing an animal down.  (You don’t have to understand the flexion reflex to make an empirical observation about effectiveness.)  Some of the stones would just bounce off the animal’s back without toppling it.  But those flat rocks with a bit of a point, given that they are spinning, will soon embed their pseudopoint into a pushed-up roll of hide.

     There’s no need to penetrate the skin.  Why does it work better at toppling the animal?  Because it stops the spinning stone, before it can bounce free.  Most throws will not strike the flank solidly, transferring all their momentum to the hapless animal.  Most will bounce off the animal’s back, slowed only slightly.  But with the pseudo-point catching the roll of hide, all of the stone’s momentum is transferred to the animal, rather than the usual fraction.

     Because of that, even palm-sized handaxe-shaped stones may be effective at toppling the animal.  The momentum transfer gets the animal moving slowly sideways, and the pain’s flexion reflex (also enhanced by the abrupt jerk when the point catches a roll of skin) interferes with the farside leg extension that usually prevents toppling.  Over it goes, ever so slowly, knocked off its feet.

     Thus the classic features of the handaxe – somewhat flattened, mostly edged, mostly symmetric, somewhat pointed – ought to be one of the nice solutions to this hominid version of the waterhole predation game.  They’d surely retrieve and reuse favorite projectiles, of course.  However, given how deep the mud is around those waterholes, they would probably lose a lot of them, too.  Over the years, quite a few handaxes would accumulate in the shoreline mud.  That’s surely a better explanation for accumulations like those at Olorgesailie than that the sites were some sort of factory, doing an early form of mass production of handaxes.

     It does suggest that the flattened teardrop is a pragmatic shape, not necessarily some innate symmetry in the mind of the toolmaker.  Even the more-of-less spherical “hammerstones” were shown by the archaeologist Nick Toth to be a consequence of repeated use, not design.  Any stone tends to become round, once you used it long enough.  As the “perfect” sphere in the mind went, so may the pragmatic teardrop.

     The Killer Frisbee use might also help explain why a few handaxes are found standing on edge in old watercourses and lake edges.  I doubt the handaxes were planted that way, just to serve as primitive sundials or to confuse latter-day archaeologists.  Thrown handaxes always land on edge, digging themselves a nice groove and often planting themselves in a vertical orientation, though most surely topple over during subsequent erosion of the landscape, as when they are tumbled by stream flow.

The layers here at Kariandusi make my point even more dramatically than the sea at Olorgesailie.  In each layer of old lakeshore silt and mud, there are hundreds of hand-sized stones showing signs of prehuman modification, most of them classic handaxes.  The next layer a few centimeters below (exposed a little farther downhill) and you see the same thing again.  And again.  It reminds me of the old jailhouse story of sending a prisoner a cake with escape tools baked inside.  Well, each layer of this “cake” is loaded with an unbelievable number of teardrop-shaped tools.

     If the waterhole predation theories of handaxe use are even partially correct, there ought to be a series of concentric rings of handaxes, layer by layer down towards present-day Lake Elmenteita, shrunken from its former self.  Well, perhaps not whole rings, as only certain sites on the lake’s perimeter would be favored by both animal trails and hominid hunting habits, but in any event there should be a lot of sites, given a million years worth of Homo erectus families to feed.

     Some Kariandusi handaxes appear somewhat smoothed, as if tumbled in flowing water – and their edges wouldn’t bite the hand that held them, making it possible to use them as a handheld axe of sorts.  But adjacent ones may have sharp edges, as if never moved downstream after being lost.  Hmmm.  Remember that floods happen, and climates eventually change.

When a river flows faster than usual, it can cut a deeper channel and so strip off layers of sediment from its banks and bottom.  New river banks appear, showing off layers of smoothed cobbles from several earlier river beds.  These are a common sight to anyone who hikes along a river today and looks at the river banks, and it would have been much the same a million years ago.  But here, some of those rock layers wouldn’t be just cobbles.  Guess what might be left exposed after all this down-cutting?

     The river banks and bottoms would contain overwhelming numbers of symmetric, pseudo-pointed rocks.  Some such “handaxes,” of course, would have been eroded out and carried downstream and tumbled smoother.  Such tool-studded river banks would become known as Handaxe Heaven in whatever language the locals had attained.  Would these early humans have called it “manna from heaven” or ascribed this bounty, perhaps, to their ancestors having provided for them?

     It lends a whole new layer of interpretation to objet trouvé.  (I can hear it now:  “No, this tribe didn’t actually make handaxes, they merely recycled the lost ones of an earlier culture.  Recycling is a million years old.”)

     And it shows you a wonderful route to reinvention.  Techniques are always getting lost, much as the Australian aborigines in Tasmania eventually lost some fire-making and tool-making techniques over the millennia, once rising sea level had isolated them from the mainland.  Cultural loss is one of the hazards of the population downsizing and fragmentation that climate change can cause.  Riverbank hunters, however, with no more than just the chimpanzeelike tendencies to lob rocks when a branch wasn’t handy, could easily rediscover the aerodynamic properties of the handaxe and become superstitious about using the “right shape.”  When the provident riverbanks were depleted, they might start reshaping local rock into the classic shape.  They’d already know, from long use, that the shape was useful.  It’s one reason why the classic shape might have survived a million years of climate change and all the island-like fragmentations of Homo erectus populations that repeatedly happened.

     Mining “handaxes” likely also occurred from sites which were no longer riverbanks, sites like Olorgesailie and Kariandusi are today.  You can consider it as either the beginnings of mining or of archaeology.  “Making use of the past” is a modern museum motto but you can see how it might have gotten started.

     There’s another kind of reuse of handaxes, too.  Consider the life history of the classic handaxe shape: if not lost first, most classic edged-all-around handaxes will eventually be broken, thanks to landing on something hard or herds trampling them.  The handaxe might lose a third of its sharp perimeter.  If thrown in the usual manner, it would tumble and not go as far, nor land on edge in the preferred manner.  But the remnant’s broken edge would be more amenable to a hand grip.  Now you could use the remaining sharp edges for all of those Swiss Army knife tasks that the archaeologists proposed in their frustration: pounding roots, defleshing, and so forth.  Ditto for the tumbled handaxes, smoothed to make them safe for the palm of your hand.  So no-one even has to give up a favorite theory.

     It’s not often when you can have your cake and eat it too.  Just remember that the handaxe has two lives, if not irretrievably lost in the lakeshore mud:  Frisbee first, Swiss second.