Aboriginal Australian weapons and human efficiency – Scientific Reports

Weapons

The kodj (kodja, kadju, käoit, coccio) and leangle (langeel, lionile, Darn-de-wan) with parrying shield (mulga, malka, elaman) (Fig. 1) used in this study were created by master tool-makers Mr. Larry Blight (kodj) and Mr. Brendan Kennedy and Mr. Trevor Kirby (leangle and parrying shield). Created for a public science series exploring Indigenous Australian weapon innovation (by Blackfella Films), these replica weapons are consistent with those used in their respective communities. The kodj was 365 mm long, 114 mm wide, and 46 mm thick. The leangle was 656 mm long, 335 mm wide, and 30 mm thick. These dimensions are comparable to known ethnographic ranges reported for the leangle^27,37 and kodj^38,39.

The leangle and shield are both made one-piece in hardwood, a raw material widely available (numerous species present) and used for creation of diverse material culture on the Australian continent⁴⁰. The kodj, on the other hand, is a composite weapon incorporating a stone axe and stone flat hammer hafted onto a wooden handle using resin³⁹. The Menang Noongar expert tool-maker of the kodj used in this study, Mr. Larry Blight, provides more insight into the components of this particular technology stating the kodj is constructed by attaching a sharpened boya (stone) blade on one side and a blunt boya edge on the other to a boorn (handle) with balga (Xanthorrhoea grass tree) resin. He specifies the resin is a combination of balga resin, kop (charcoal), and yonga (dried kangaroo faeces), which when combined is known as biriny and is coated over the boya and boorn while hot and plastic before drying to a hard binding.

Both weapons were used to strike at an opponent. The leangle is used in hand-to-hand combat and designed such “that the warrior may be enabled to strike round the shield, or elaman, of his adversary”⁴¹ with Etheridge³⁷ reporting that “when a combatant wishes to strike side-wise and from himself with a back-handed blow, the round, and not the point, of the Leonile, is used”. Brough Smyth⁴² describes the langeel as “perhaps the most dangerous of all the weapons of [hand-held offensive weapons]…because of the facility with which the point can be suddenly turned at the moment of striking, [and] is [thus] difficult to avoid”. The mulga or parrying shield which accompanies the leangle was designed to be narrow (maximising the warrior’s visibility of his opponent) and bow out to a thick convex centre to better parry away blows from the opponent. These shields often featured engraved and painted designed reflecting the community which made the weapon.

Less is reported ethnographically regarding the use of the kodj in fighting, though, like the leangle, it can be “pivoted by a turn of the wrist so that the blade can cut in any direction”³⁸. Breton⁴¹ describes this tool as one used to cut notches in tree trunks so that hunters could ascend into the high branches and adds that he has “no doubt [they] use them in their wars as well”. Similar early descriptions of its use in hunting are found in King’s observations of the people of King George III Sound, where he describes the käoit as being used in “killing seals and other animals by striking them on the head”⁴³, while Salvado (quoted in³⁹) stated that the Kodj was used “to secure game from dead trees, cut footholds on trunks of trees to enable them to climb, fashion their weapons, break the bones of the kangaroo, and other animals in order to extract the marrow for food or to anoint themselves, and a thousand other uses”.

A multi-view three-dimensional digital reconstruction of each weapon was created from a high definition cyberscan (Myriad Studios; myriadstudios.com.au) in Fusion 360 software (Autodesk, CA, USA). The leangle was carved from solid hardwood and assumed to have a constant density (1.130 g.cm⁻³), computed from measured mass (1.055 kg) and mesh volume (933.716 cm³). The kodj density (1.148 g.cm⁻³) was calculated from total measured mass (0.350 kg) and mesh volume (304.949 cm³). To better approximate centre of mass and moments of inertia, the kodj handle and head densities were calculated independently. Handle mass was calculated using the mesh volume of the extruded handle and the density of dried acacia/wattle (570 g.cm⁻³)⁴⁴. The density of the kodj head was assumed constant and calculated from the remaining mass (measured mass − handle mass) and the mesh volume of the head. Centre of mass and moments of inertia (xy, xz, yz) of each weapon were subsequently computed using Fusion 360 proprietary tools.

Biomechanics

Three-dimensional motion of the upper body joints during a strike, the corresponding moments of force acting about these joints, and subsequent powers produced by these joints can be quantified using data obtained from inertial measurement units (IMU). Integrating these data into a neuromusculoskeletal model of the human body⁴⁵, with consideration of the physical properties of the weapon (e.g., centre of mass, inertia), enables a comprehensive analysis of the human biomechanics involved in striking using the kodj and leangle.

Biomechanical data were collected from one adult male Aboriginal Australian who used both weapons. The participant had prior experience using the weapons but did not consider himself an ‘expert’. Ethical approval was obtained from the Griffith University Human Research Ethics Committee (GU#2022/665). The participant was informed of the procedures and provided their written informed consent, consistent with the Declaration of Helsinki, prior to participation.

Seventeen IMU (Xsens/Movella, NV, USA) were attached to the participant’s body. The IMU were placed according to manufacturer guidelines on the head, shoulders, upper arms, forearms, hands, torso, pelvis, thighs, shanks, and feet. Twenty-eight retro-reflective markers (SI Fig. 1) were placed atop bony landmarks and a static trial (i.e., quiet upright stance) of three-dimensional marker locations was record at 200 Hz in Nexus 2.12 (Vicon, Oxford, UK). The participant performed static and functional calibrations to initialise the orientation of the IMU. The leangle was used with the parrying shield, and the participant performed seven repeated trials of a striking motion using the leangle and then the kodj (Fig. 5). All IMU data were recorded at 100 Hz using proprietary software (MVN Analyze, Xsens).

To analyse body kinematics and kinetics, a custom model was constructed in OpenSim version 4.4⁴⁵. A bilateral upper extremity model⁴⁶ with glenohumeral, scapulothoracic, and lumbar motions as well as 35 degrees of freedom was used as a base. The lower extremities were added from a previously validated model⁴⁷, as was a three-axis neck joint. The weapons were attached to the hands of the model and aligned using video recorded concurrent and simultaneous to IMU data. The three-dimensional marker locations were used to linearly scale model dimensions and inertial properties in OpenSim.

The orientations of each body segment and the translations of the pelvis from the reconstructed motion in MVN Analyze were exported and read in MATLAB version 2023a (Mathworks, MA, USA). The orientations of each body segment and the translations of the pelvis were applied to the OpenSim model using inverse kinematics tool with equal tracking weights for all segments of the body. Joint moments were computed with inverse dynamics using OpenSim inverse dynamics tool, during which the input coordinates were low pass filtered with a 6 Hz zero-lag dual-pass 3rd order Butterworth filter. Joint powers were calculated as the product of joint moments and joint angular velocity.

For each trial, the centre of mass velocity was calculated for each segment of the body (e.g., thigh, shank, forearm, head) and weapon in OpenSim. Whole body centre of mass velocity was calculated using a weighted sum, according to:

where m_i is the mass of a segment of the body, v_i is the velocity of a segment of the body, n is the number of body segments, and m_T is the total mass of all body segments and weapon(s). Kinetic energy was calculated for each weapon and segment of the body according to:

where KE is kinetic energy, m is the mass, v is the velocity, I is the moment of inertia, and ω is the angular velocity. Combined person and weapon (total) kinetic energy was calculated as the sum of all segments of the body and weapon(s). The proportion of weapon kinetic energy with respect to total kinetic energy was also calculated at the timepoint when peak weapon kinetic energy occurred. After calculating joint angles and moments, centre of mass velocity, and kinetic energy, each strike was time-normalised to 101 points to allow for ensemble averaging and between-weapon comparisons. The start of the strike was identified as the timepoint during which the weapon centre of mass was most posterior to the participant’s centre of mass during the backswing. The end of the strike was identified as the first timepoint in which the weapon centre of mass was lower than the elbow.