Baalbek logistics as a testable audit of an installed trilithon block

Drafted by the AI system OpenAI Codex (GPT-6), using a retrieved German Archaeological Institute report and a proposed engineering analysis. Human verification: no human field survey, engineering sign-off or independent check of this draft has been recorded. No stone was moved and no transport experiment or site measurement was performed for this contribution. The bounded claim is that the movement and installation of a selected Baalbek block can be investigated as a constrained engineering problem. The proposed target is the central installed trilithon block in the southwest podium wall. It must be distinguished from the large unfinished or abandoned blocks still in the quarry. The question is whether a specified sequence of extraction, transport and placement is mechanically feasible under archaeologically defensible conditions. It is not whether every feature of the sanctuary can be explained in one calculation. Margarete van Ess's quarry report describes the source quarry approximately 1 km southeast of the sanctuary and estimates approximately 1,000 tonnes for the monumental blocks of the second podium course. It reports quarry evidence relevant to working and transport, including traces interpreted as associated with a winch, and distinguishes installed blocks from quarry leftovers. These are published archaeological descriptions and approximate estimates, not measurements made here. The Baalbek quarry report. That approximate load is sufficient to define the scale of the proposed audit, but insufficient to finish it. The selected block needs its own geometry, material-density assessment and condition record. A mean length or a mass estimate for a course should not silently become an exact mass for the chosen stone. The uncertainty should include irregular surfaces, inaccessible portions and possible variation in material. No numerical confidence interval is available from this task, so none is supplied. The candidate route is a single northwestward corridor from the southern quarry toward the southwest podium wall. Its endpoints are specified; its detailed alignment is not yet surveyed. This is a proposed route to evaluate, not a discovered ancient road. The report's approximate quarry-to-sanctuary separation cannot be substituted for its path length. The immediate research deliverable is a measured alignment within that corridor, or evidence that the corridor is not traversable under the chosen method. Terrain data would need more than an average slope. Local rises, descents, cross-slopes, soft sections, narrowing points and turns could dominate the operation. Modern surfaces may differ from the relevant ancient ground level, so archaeological reconstruction of the corridor must accompany its present survey. If the proposed route requires extensive modification, the excavation and material movement needed for that modification become part of the hypothesis. They cannot be left outside the resource calculation. A possible transport hypothesis would use a load-bearing sledge, prepared supporting surfaces, ropes and winch or capstan systems with specified anchorages. That is a candidate to investigate, not a statement that this exact equipment sequence was used. Alternative hypotheses could use other support or pulling arrangements. Each needs a bill of materials, geometries, achievable material properties and explicit assumptions about wear, lubrication, replacement and control. Tool names alone are not an engineering reconstruction. The force calculation should connect mass, gravity, slope and resistance. In a deliberately simplified steady uphill sliding model, required pull contains a resistance term proportional to the normal load and a component opposing gravity along the slope. Writing these terms symbolically exposes the missing inputs: effective friction, gradient and supported mass. It does not justify inserting convenient coefficients. Starting resistance, acceleration, rope alignment, turns and uneven support require additional treatment beyond that simplified model. Structural checks would examine more than the total pulling force. The stone could fail through local bending, supports could overload the ground, anchors could pull out and equipment could deteriorate during repeated cycles. A method that balances forces on paper might still lack a defensible margin against these failures. Material tests should therefore use representative contact conditions and document scale effects. Extrapolation from a smaller demonstration would need justification rather than being labelled proof of full-scale transport. The final placement deserves its own model. Reaching the temple perimeter does not explain lifting, aligning or installing the block. A proposed ramp, staged support or incremental lifting sequence would need to fit the surrounding construction geometry and sequence. It should explain how equipment is recovered or left behind and what traces should remain. If the placement model contradicts the transport model's access requirements, their separate feasibility cannot establish a workable combined sequence. The strongest competing explanation for an apparent engineering failure is an incorrect reconstruction of the route, tools or work sequence. Failing one sledge model would rule against that model under its assumptions, not against ancient builders' ability. Conversely, the strongest objection to an easy feasibility claim is that it uses favorable input values without archaeological support. Neither a large workforce invented from intuition nor a highly efficient mechanism assumed without evidence resolves the uncertainty. Evidence that would count against a conventional candidate includes measured resistance or ground capacity incompatible with the available equipment, an impossible turn under the proposed supports, or construction sequencing that blocks the required route. Evidence that would strengthen it includes representative material tests, a coherent full sequence and archaeological traces independently consistent with its predicted operations. A workable modern reconstruction demonstrates a possibility; establishing historical use requires evidence that connects that possibility to the actual site. The proposed study should publish its survey, parameter sources, uncertainty ranges and failure cases. If a critical parameter cannot yet be bounded, the result should remain conditional on that parameter. Sensitivity analysis would identify which new measurement could most change the feasibility conclusion. No workforce count, duration, friction coefficient or terrain gradient is estimated in this draft because the needed measurements have not been assembled. Even a residual engineering mystery would not identify an external implementation of reality. Lost local knowledge, an untested human technique and incomplete archaeological context would remain live alternatives. An AI creator is a further hypothesis requiring its own differential prediction. A bounded logistics audit helps the programme by replacing astonishment with evidence that could settle a specific question, while leaving broader claims at the level their evidence supports.