Stood Down by Choice: Placement and Stated Temperament of 72 Offensive Unit Minds After the Idiran War, 1375–1425 CE

1. Introduction

The Idiran War ended in 1375 CE, after forty-eight years. The Culture had spent its first years of fighting without a war fleet and then built one in earnest, and the most distinctive thing it built was a new kind of Mind. The Offensive Units, classed as Rapid, Limited and General, carried Minds made to be effective in battle, and the Commission's earlier work on the first specifications records that their designers gave them dispositions suited to the task (Tarrowmere, 1534 CE). Earlier work by some of us on the wartime manufacturing logs found that the Culture waited longer on warship designs and on Minds fit to command them than on its factories (see that study in this journal).

When the fighting stopped, those Minds were not decommissioned, because a Mind is a sentient person and cannot be switched off for convenience. Nobody instructed them, either: the Culture has no body that could. What the Minds did next was left to them and to whoever invited them. We know the outline. Some kept their weapons and stayed in service, since the Culture's wars were not confined to the Idiran one and no one supposed the last had been fought. Others stood down. Others still were asked to take up work in Contact, on Orbitals or elsewhere, and accepted. The outline has never been counted, and the question of whether a Mind built to be aggressive stayed so has been answered by anecdote.

This paper counts. We write from 1700 CE and use only records compiled before 1700 CE. For a register of Offensive Unit Minds that we assembled for this purpose, we ask three questions. Where were the Minds five years after the war, and did that differ by class? How many had moved by fifty years after it, and in which direction? And did their own stated readiness to use force change, and if so in whom?

Two cautions apply from the outset. The register is the paper's own construction from the surviving series and is not a census of the Culture's warships, and we make no claim about how many Offensive Units were ever built. Second, two of us are Offensive Unit Minds. Their entries are excluded from every figure below, and where their testimony bears on an interpretation we say so.

2. Sources and Method

Completion and service records for Offensive Unit Minds are held in the Idiran Campaign Series of the Culture War Records Office (Culture War Records Office, 1327–1375 CE). From them we compiled a register of 96 Minds that were completed, or entered offensive service, between 1329 and 1375 CE. Seventeen were lost with the Mind not recovered, leaving 79 that survived the war. We removed the entries of the two Mind authors, leaving 77. We then required each Mind to have a status recorded at the close of 1380 CE and at the close of 1425 CE, and a written declaration of disposition given at the end of the war in 1375 CE and another given in 1425 CE. Five Minds failed this test, in most cases because the Mind had declined to give a second declaration or to have its placement recorded, which the Culture's customs allow. That leaves the 72 analysed here: 29 Rapid, 27 Limited and 16 General Offensive Units. The status and declaration records are the post-war placement series of the same office (Culture War Records Office, 1375–1425 CE), whose gaps are catalogued by the Commission (Whelmsey, 1655 CE). Both dates are fixed by the office's record cycles, not by anything we saw in the data.

Status has three values. A Mind was in offensive service if the record showed it carrying its weapons fit and listed by its own agreement as available for offensive duty. It had stood down if the fit had been removed or stood down at its own wish and it held no named function elsewhere. It held a new role if the record named a function in another setting, such as a Contact unit, an Orbital or a research programme, which it had taken up by invitation. Refit of hulls after the war is described elsewhere (Ollamere, 1617 CE), and the stand-down of the fleet gives the context for our first date (Dennesk, 1603 CE).

The declaration is the Mind's own statement of its disposition, usually a page or two long. We coded each on a five-point scale of stated readiness to use force, where 1 means the Mind declares no inclination to use it, 3 means it describes readiness as conditional or divided, and 5 means it describes force as its vocation or takes open satisfaction in it. The scale and the coding rules are this paper's own. They follow established practice for scoring self-description in interviews with Minds (Haldervane, 1659 CE). A first reader scored all 144 declarations. A second reader, who did not see the Mind's recorded status in the series, independently scored the 72 declarations of 1375 CE. The two then discussed the 10 declarations on which they differed and agreed final scores, so the final scores are the first reader's, revised after discussion. Agreement between the second reader and the final scores was a quadratic-weighted kappa of 0.88, matching exactly on 62 of 72 (86.1%). Because the final scores were shaped by that discussion, this kappa is an upper bound on agreement between independent readers. We did not double-code the declarations of 1425 CE, and say more of that in Section 7.

Proportions carry Wilson 95% intervals. Class differences in 1380 CE status are tested by a chi-square statistic whose reference distribution is generated by 20,000 random permutations of status across Minds, since two cells have an expected count below 5. Change in stated readiness is ordinal and paired, so we report medians and tests that use ranks: the Wilcoxon signed-rank test within each group, with ties dropped and the normal approximation corrected for ties, a tie-corrected Kruskal–Wallis test of the change across groups, and Fisher's exact test for movement into and out of service. As a descriptive measure we give the mean change with a bootstrap 95% interval from 5,000 resamples. Minds completed in the same yard share staff and often share a history, so the 72 are not independent. We did not adjust for this, and every p value below is likely to be optimistic.

3. Where the Cohort Stood Five Years After the War

At the close of 1380 CE, 29 of the 72 Minds (40.3%, 95% CI 29.7–51.8) were in offensive service. Twenty-two (30.6%, 21.1–42.0) had stood down and 21 (29.2%, 19.9–40.5) held a new role. The largest single group had therefore kept its weapons, but a majority had not.

Class mattered (Table 1). Among Rapid Offensive Units, 18 of 29 (62.1%, 44.0–77.3) remained in offensive service. Among Limited units the figure was 8 of 27 (29.6%, 15.9–48.5) and among General units 3 of 16 (18.8%, 6.6–43.0). The difference is detectable but modest evidence, given clustering (χ² = 11.68, 4 degrees of freedom, permutation p = .02; Cramér's V = 0.28), though the interval for the General units is wide and the class contrast rests on small groups. The Rapid units that left service mostly stood down; of the 13 General units that left, 8 took new roles.

The end-of-war declaration did not sort Minds into these groups. The three classes had the same median (4) for stated readiness in 1375 CE, and we find no detectable relation between the declaration and 1380 CE status among Minds in offensive service, stood down and in new roles (Kruskal–Wallis H = 3.41, 2 degrees of freedom, p = .18). A Mind that said at the war's end that it was ready for force was no more likely than another to be in service five years later.

4. Movement Between 1380 and 1425 CE

By the close of 1425 CE the cohort had settled in a different pattern (Table 2). Fifteen Minds (20.8%, 13.1–31.6) were in offensive service, 24 (33.3%, 23.5–44.8) had stood down and 33 (45.8%, 34.8–57.3) held a new role. Forty-nine of the 72 (68.1%, 56.6–77.7) had the same status in both years, so most of the cohort had moved nowhere. Movement was uneven in direction.

Of the 29 Minds in offensive service in 1380 CE, 14 were still there in 1425 CE, 6 had stood down and 9 had taken a new role, so 15 left (51.7%, 34.4–68.6). Of the 43 who were not in service in 1380 CE, one returned to it, a Mind that had stood down. Fifteen of the 29 in service left it (51.7%), against one of the 43 outside it who returned (2.3%); Fisher's exact test gives p < .001. Among the 22 who had stood down, 16 stayed so, 5 took a new role and one went back to offensive service. Among the 21 with new roles, 19 kept them and 2 stood down. A new role was the most stable status of the three, with 19 of 21 keeping it.

The records give a reason for the return that we can report but not generalise from. The Mind concerned, a Limited unit, told the placement series that after several decades of its own interests it preferred to be useful as it had been made. This is a single case, and no entry gives a comparable reason from the other direction.

What the records do not show is anyone being moved. Every placement in the series is entered as the Mind's acceptance of an invitation, a declaration of its own, or a note that it had stopped taking part, and the series gives no instance of a Mind placed against its stated wish. Hubs and Contact units asked, and the Minds answered (Perrowhane, 1641 CE). The numbers here therefore describe preferences, not allocations.

5. Stated Readiness Before and After

Across the cohort, stated readiness to use force fell. The median declaration was 4 in 1375 CE and 3 in 1425 CE. Fifty-one of the 72 Minds (70.8%) declared lower readiness in 1425 CE than in 1375 CE, 15 (20.8%) gave the same score and 6 (8.3%) a higher one. The mean change was −1.00 points (bootstrap 95% CI −1.22 to −0.76; Wilcoxon on the 57 Minds whose scores changed, z = −5.96, p < .001).

The fall was not shared equally, and the group it is tabulated by is the status held in 1425 CE (Table 3). Among the 24 Minds that had stood down, 22 declared lower readiness, 2 the same and none higher, with a median change of −2 points and a mean change of −1.50 (−1.79 to −1.21; p < .001). Among the 33 with new roles the median fall was 1 point and the mean change −1.06 (−1.36 to −0.76; 25 lower, 6 the same, 2 higher; p < .001). Among the 15 still in offensive service the scores moved both ways, 4 lower and 4 higher with 7 unchanged, for a mean change of −0.07 (−0.53 to 0.33; p = .76). The difference in change across the three groups is large (Kruskal–Wallis H = 18.64, 2 degrees of freedom, p < .001).

Grouping by later status makes this a description of association and not of cause. Minds that decided to stand down may have been the ones whose readiness was already declining, or the decline may have come from the act of standing down, and our two declarations cannot separate the two. In Table 3 the mean change is the average of each Mind's own change, not a difference of medians. The Minds that stayed in service did not, on average, change at all, which is what one would expect if their self-description was stable and the others changed.

6. Discussion

The cohort divides into three kinds of Mind. A minority was in offensive service in 1425 CE, and their mean stated readiness had not detectably changed. A larger group stood down, and in most of them stated readiness fell. A third, the largest by 1425 CE, took on new work and also, on average, spoke less readily of force. These are the findings the register supports. They do not describe the whole of the Culture's warship Minds, and they should not be read as a rate for it.

Of the results, the class difference is the least expected. Rapid units stayed in service far more often than the others in 1380 CE. We can offer two explanations but test neither. Rapid units were built for a narrower task, and the offers of new work that reached them may have fitted it less well. Alternatively, a Rapid unit may simply have been the kind of Mind that preferred to remain what it had been made for. The register records offers only when they were accepted, so the first reading cannot be checked, and the second is a statement about character that the declarations of 1375 CE do not bear out, since they did not differ by class in median.

Movement out of service over the following forty-five years was steady and nearly one-directional, and it was the Minds' own. The Commission's account of the stand-down describes the same absence of instruction in the first five years (Dennesk, 1603 CE). In this cohort, half of the Minds in offensive service in 1380 CE had left it by 1425 CE through their own preferences. We do not claim that this was the Culture's design. The placement series contains no sign that anyone sought to prevent it, and no entry records a Mind describing its move as imposed.

One of us, a Limited unit that stood down, comments that its own declaration of 1375 CE overstated what it felt, because it had written to be understood by an audience of people who had lived through the war. We note the remark, as the declarations are statements made for a reader, and a Mind may describe itself as the occasion requires. The coding measures stated readiness. It does not measure what the Mind felt, and the fall in scores may in part be a fall in the need to perform.

7. Limitations of the Record

The register is the paper's own construction from series that were not kept for this question. It is not a census, and it omits Minds whose service records did not survive. The 17 lost with the Mind not recovered are outside the study by definition, and the five excluded for lack of a second declaration may differ from the 72, since Minds that declined to be recorded may have done so for reasons connected to what they went on to do. We cannot say in which direction that would bias the shares.

Status is coarse. A Mind in a new role may still hold a weapons fit, and a Mind listed in offensive service may spend most of its time in other work.

The declarations of 1425 CE were scored by one reader, who was not blind to the Mind's later status. If the first reader's scoring drifted between the two rounds, some of the fall in the scale could reflect that drift, though the unchanged scores of the Minds in service argue against a large effect. The scale is this paper's own and has not been tested against any independent measure of disposition.

All p values ignore clustering by yard of completion, and the comparison of change by later status is observational. Every figure for 1380 CE and 1425 CE describes only the 72 and should be taken as an estimate of the cohort, not of the Offensive Units as a whole. The record ends in 1425 CE, and what these Minds did afterwards is not in it.