Manufactured Ferocity: Selection, Shaping and the Retention of Sardaukar Training on Salusa Secundus, 9984–10191 AG

1. Introduction

For most of the Corrino era the origin of the Sardaukar was a guarded secret of the Padishah Emperors. The Landsraad knew the Emperor's legions as the finest fighting force in the Imperium, and knew less about where they were raised. The secret was broken in 10191 AG, when Paul Atreides confronted Shaddam IV at Arrakeen and used his private knowledge that the legions came from the prison planet as leverage in the settlement that ended Shaddam's reign; it came into scholarly view only once the Corrino archives were released. Shaddam then went into exile on Salusa Secundus itself. We write c. 10242 AG, some decades after the Jihad of Muad'Dib, when Corrino internal records have been open to scholars for a generation.

Salusa Secundus did not begin as a penal world. It was the seat of the Imperial court until a catastrophe, whose nature the surviving record leaves obscure, devastated the planet; the court moved to Kaitain, and only then was Salusa designated the Imperial prison planet (Varrenkamp, 10219 AG). The hostility of the planet is therefore a product of that catastrophe and of centuries of use as a place of detention. What requires explanation is why the Corrino emperors, who could have trained their legions anywhere, kept doing so on a ruined world.

Two mechanisms could make the environment useful. Under selection, the planet kills the weak before any instructor sees them, and recruiters take the survivors. Under shaping, surviving the planet changes the survivors, so that a recruit raised and trained there becomes a better soldier than an identical recruit trained elsewhere. We reconstruct the evidence for each and then ask what the court's own deliberations reveal about its understanding of them.

2. Sources and Method

Three archival series carry the argument. The first is the relocation file of the Office of the Imperial Household Bashar: fourteen memoranda attached to five formal proposals, 9984–10190 AG, captured at Kaitain after 10191 AG and released under the Atreides administration (Office of the Imperial Household Bashar, 9984–10190 AG). A bashar is a senior Sardaukar field officer, ranking somewhat above a colonel; the household bashar advised the Emperor on the legions. We coded each memorandum for the grounds it gave for or against relocation. A memorandum could cite several grounds.

The second series is the detention-district birth and mortality registers kept by the Salusa Secundus Penal Administration (10120–10177 AG). They survive for 23 birth cohorts in four districts, covering 4,860 children born in 10120–10165 AG and followed to recruitment age. The third is the intake and re-examination returns of the Sardaukar Recruitment Office for a trial conducted in 10143–10149 AG (Sardaukar Recruitment Office, 10143–10150 AG), described in Section 3.

Secondary evidence comes from curricula of the Salusan field camps (Kesteven, 10231 AG), Mentat threat appraisals of Sardaukar forces (Aldoran, 10228 AG), Harkonnen correspondence of 10190–10191 AG, and Fremen testimony on the battle of Arrakeen. All three primary series were compiled by the institution whose policy they record, and the recruitment returns in particular were written by an office with an interest in the outcome. We report proportions with the range across cohorts, estimate the trial comparison with a linear mixed model, and treat every figure as approximate.

3. Selection and Shaping on Salusa Secundus

Selection on Salusa begins in childhood, across the whole detained population, and long before recruitment. Of the 4,860 children in the registers, 2,190 (45%) died before the age of eleven. The proportion ranged from about 38% to 53% across cohorts, and we would not defend a figure more precise than "approaching half". Recruiters screened survivors at eleven or twelve. Of the 2,670 survivors, 612 (23%) were taken into the legions. The Sardaukar were thus drawn from roughly one child in eight of those born, after two filters, one environmental and one administrative.

Such figures cannot by themselves show shaping, because every survivor passed the same filter. The trial of 10143–10149 AG offers a rare lever. Following the fourth relocation proposal, the Recruitment Office sent 96 recruits from the 10143 AG intake to an enclosed training installation in the Kaitain system, and kept 214 recruits from the same intake in the Salusan field camps. The returns state that allocation was by lot, a claim we cannot verify independently. Both groups had therefore survived the same childhood and passed the same screening; any later difference between them should reflect training conditions.

The measure recorded at intake and at the six-year re-examination was load-march endurance, the hours a recruit could sustain a standard loaded march before collapse or withdrawal. At intake the groups were close (Salusan minus off-world difference 0.2 hours, 95% CI −0.4 to 0.8, p = .52). A linear mixed model with a random intercept for each recruit gave a group-by-time interaction of 3.1 hours (95% CI 1.9–4.3, p < .001): Salusan recruits gained about 5.4 hours over the period, the off-world cohort about 2.3.

That estimate is itself vulnerable to selection. Of the 214 Salusan recruits, 41 (19%) died or were invalided out before re-examination, against 7 of 96 (7%) off-world (χ²(1) = 7.1, p = .008; odds ratio 3.0, 95% CI 1.3–7.0). The Salusan camps continued to remove weaker recruits during training. When we assigned each lost recruit the lowest re-examination value recorded in his group, the interaction fell to 2.2 hours (95% CI 0.9–3.5, p < .001). Because this worst-case imputation treats every loss as the weakest possible outcome, 2.2 hours is best read as a lower bound on shaping. Up to about a third of the raw advantage (0.9 of 3.1 hours) may reflect continued selection; the true share could be smaller.

4. The Instructional Layer

Environmental hardening was never the whole of Sardaukar training. Camp curricula of the later Corrino period devote most drill hours to bladework, small-unit movement and close assault (Kesteven, 10231 AG). The emphasis on the blade follows from shield physics: because the Holtzman shield stops fast objects and admits a slow one, projectile weapons are of little use against a shielded enemy, and the decisive skill is the controlled slow thrust. That instruction was portable, and the off-world cohort received it in full.

Command structure is the second instructional element. The legions were commanded by a burseg, a general officer, with detachments under bashars and a levenbrech, a junior officer, serving under each bashar. Sardaukar discipline was proverbially rank-proud and obedient. Drill in that hierarchy, like bladework, could be taught anywhere, and the returns record that the off-world cohort was organised and exercised under the same ranks (Sardaukar Recruitment Office, 10143–10150 AG). Whatever the Salusan camps added beyond the off-world installation, then, it was not a different chain of command.

Performance under Shaddam IV is harder to document than reputation. The best-recorded action is the covert Sardaukar assault on the Atreides at Arrakeen in 10191 AG, fought in Harkonnen uniform, which destroyed the ducal household's forces. Mentat appraisals from Atreides service rated a single Sardaukar legion as a match for the levies of most Great Houses (Aldoran, 10228 AG). The legions were also the coercive backstop of Corrino rule, the force behind the residual appeal to arms in Imperial succession, which helps explain why anything touching their production was guarded so closely.

5. Relocation Deliberations

Five relocation proposals reached the Emperor's council in the period (Table 1). Each arose from cost or from the difficulty of supplying camps on a devastated world. Each was rejected, the fourth only after the trial of Section 3 had run its course.

The grounds for rejection cluster tightly. Environmental hardening appears in 11 of 14 memoranda (79%), generally as the claim that recruits trained elsewhere would be softer. Secrecy appears in 9 (64%): a legion raised on another world would reveal that it was being raised, and the Landsraad might ask where its predecessors came from. Cost appears in 4 memoranda (29%) and throne security in 3 (21%). The memorandum closing the fourth proposal in 10150 AG cites the trial returns directly and recommends that no off-world cohort be raised again.

We read the file as showing tacit knowledge of the environment's contribution, and increasingly explicit knowledge after 10149 AG. The court did not describe selection and shaping in those terms, but it repeatedly accepted the costs of a ruined world on the ground that the ruin was doing work no instructor could replace.

6. Discussion

If hardship is an input to military production, a harsher world should produce a better force. Arrakis supplied the test. Thufir Hawat, then in Harkonnen service, and Paul Atreides each recognised the desert planet as a harder Salusa Secundus. Harkonnen correspondence of 10190–10191 AG shows the Baron contemplating a deliberate version of the same process: oppression of Arrakis under Glossu Rabban, intended to harden a population from which a force of Sardaukar quality might later be drawn (Office of the Baron, 10190–10191 AG). The plan treated the Salusan effect as reproducible by policy.

The Fremen had been undergoing that process for generations without Harkonnen design. Their mobility advantage over Harkonnen combined-arms forces before 10191 AG is well documented, and in 10191 AG, riding worms through a storm into the Arrakeen basin, Fremen fighters defeated the Sardaukar escorting the Emperor. Fremen testimony credits the Fedaykin, the death commandos of Paul's following, with the heaviest fighting (Sietch Tabr Oral History Collection, 10196–10240 AG). On the day, the force raised on the harsher world prevailed.

Heavy qualifications apply. The Fremen had surprise, the storm, worm transport and atomics against the Shield Wall, none of which is a product of hardship, and a single confounded battle cannot isolate the effect of upbringing. Arrakeen is therefore corroboration at most; the causal weight of our argument rests on the trial of 10143–10149 AG. Paul's subsequent choice of Fremen as the shock troops of his Imperium repeated the Corrino logic on a larger scale. The trial, the council's reasoning and the Arrakeen outcome all point the same way: the environment selected and shaped, and those who controlled such worlds understood it.

7. Limitations of the Record

Every primary series was written by an interested party. The relocation file may overstate the unanimity of the council, since dissenting memoranda may not have been filed with the proposals. Mortality registers survive for four districts, which may not represent the planet. The trial rests on one intake, one physical measure and an allocation by lot that the recruitment returns assert but do not document. Our worst-case treatment of training deaths bounds continued selection without measuring it, so the split between selection and shaping within the 3.1-hour advantage remains uncertain.

Sardaukar production after 10191 AG, under the exiled court and Atreides supervision, lies outside this study. The record for that period is thin, and we draw no conclusion from it.