3

The psychology of strategic decision-making: Ambon 1941–42

Joan Beaumont

I first met David Horner in the Public Record Office (now The National Archives) in London. The Public Record Office had recently moved to Kew, and David and I were queuing to collect documents or order photocopies at an unnerving price that could only be charged in the pre-digital age. It was 1977, and we were both aspiring postdoctoral scholars in our late twenties. David told me of his forthcoming book, Crisis of Command: Australian Generalship and the Japanese Threat.¹ Little did I know that this would be the first of (at last count) 38 books he would go on to write, co-author and edit. Crisis of Command explored a theme that would infuse much of Horner’s later work: how well did Australia’s political and military leaders respond to threats to the nation’s security?

The years 1941 to 1942 provide a classic case study of this issue, since, to quote Horner, this was the moment in Australian military history when the Australian high command was subject to ‘the greatest stress’.² In the frantic three months after the outbreak of war with Japan, ‘[t]he strategic situation deteriorated rapidly and the threat to Australia mounted daily’.³ When judging how well Australia’s military and political leaders responded to this crisis, Horner’s conclusions were judicious and balanced—qualities that characterise all his work. While Australian Army leaders failed to anticipate the speed of the Japanese advance, Horner concludes that they could not be held accountable for the dire strategic situation confronting Australia. The deficiencies in the organisation of Army Headquarters for command, the lack of trained troops in Australia, the shortage of equipment and the dearth of young experienced leaders were undeniable, but ‘[p]ractically none of these shortcomings should be laid at the feet of the army leaders’.⁴ Given the fiscal and industrial constraints on rearmament in the 1930s, and the agreement of the government of Robert Menzies to contribute the majority of Australian resources to the war in Europe and Malaya, the naval and air forces in Australia in the latter part of 1941 were ‘almost non-existent’.⁵ Hence, the options for defending Australia’s immediate north against Japan were limited. Within the context of these constraints, Horner concludes that although Australia’s military leaders made mistakes, ‘and their conduct was not always as pure as it might have been, generally speaking Australia was well served by her generals in 1942’.⁶ Specifically, he praises the Chief of the General Staff, Lieutenant General Vernon Sturdee, who brought to crisis management his sense of propriety, a recognition of the importance of sticking to the rules and ‘his determination to face the disasters in a logical and calm manner’.⁷

There were, however, exceptions that might qualify this judgement. Among them was the decision by Sturdee in January 1942 to leave an Australian force consisting of the 2/21st Battalion and ancillary troops on the island of Ambon, even when its commanding officer insisted that it was doomed to fail and should be evacuated. Horner concedes that the deployment of Gull Force and its equally unfortunately named counterparts, Sparrow and Lark Forces, to Dutch Timor and New Britain respectively was ill-conceived: it was a ‘token contribution to forward defence … merely grasping at straws’. Army Headquarters must be ‘indicted for failing to assess realistically the chances of these garrisons’. Horner attributes the fate of these garrisons—all of which were rapidly defeated in early 1942—to the fact that the army leaders found themselves caught between the cleft stick of Singapore and the defence of Australia. ‘If Singapore held firm with a substantial naval force ready to strike at the Japanese, then there was little need to strengthen the garrisons in the islands.’⁸

This chapter does not seek to challenge this assessment but suggests that we need to understand more about the processes of decision-making concerning the islands strategy. Notably, three questions invite further exploration. First, why was the initial decision to support the Netherlands East Indies (NEI) taken with so little rational analysis of the possibility of successful defence? Second, why were Australian forces still sent to the NEI when it was clear, after the sinking of the Royal Navy’s HMS Prince of Wales and HMS Repulse on 10 December 1941, that Singapore had no naval defences, and that the NEI might become the first line of defence? Finally, why, when the commander on the ground predicted disaster, were Australians left on Ambon to fight a hopeless cause rather than be withdrawn and concentrated elsewhere: perhaps in Malaya, northern Australia, or even the north coast of Papua, where the Japanese attacked in July 1942?

The answers to these questions lie partly in the documentary record of which Horner is such a master, but arguably we can only fully explain this ‘penny packet’ strategy if we delve into the psychology of decision-making. This is inherently perilous terrain for military historians, and one that has been neglected in Australian historiography. However, the decisions made in the crisis of 1941–42 were not always rational. Rather, they were shaped by emotions, intuitive preferences and biases of which the decision-makers themselves were possibly unaware. Horner acknowledges the importance of emotions: ‘Men’ he writes in Crisis of Command, ‘are not automatons and their decisions are governed as much by emotion as logic.’ These emotions, he continues, can be discerned in part by reading the commanders’ personal correspondence, wherein is revealed ‘elation and depression, suspicion and confidence, frustration and satisfaction, fortitude and trepidation’.⁹

However, this chapter argues that emotions are only one element in what might be called modes of cognition. In arguing thus, I draw on the analysis of decision-making by the late Nobel prize–winning Israeli-American behavioural psychologist Daniel Kahneman. In his critically acclaimed 2011 bestseller, Thinking, Fast and Slow, Kahneman argues that the machinery of human cognition incorporates two modes of thought: ‘System 1’, which is fast, instinctive, automatic and emotional; and ‘System 2’, which is slower, more deliberative, and more logical, but arguably less persuasive.¹⁰ The catastrophe of Ambon almost defies explanation unless we see it as an instance in which System 1 infused, even dominated, System 2 in the thinking of Australia’s military leaders.

Kahneman’s analysis is preferred here to the classic 1976 study of military incompetence by the experimental psychologist Norman Dixon, although this, too, provides useful insights.¹¹ Dixon’s four key symptoms of military incompetence—wastage of manpower, overconfidence, underestimation of the enemy and the ignoring of intelligence reports—were, as we shall see, manifested by Australian military leaders in 1941–42. However, Dixon’s ultimate conclusion, following his analysis of mostly British military disasters over the last century and half, is that military incompetence was not a function of stupidity but rather of the authoritarian personality of commanders and of the military systems that attracted these types. Some non-authoritarian generals might also have been hopelessly incompetent, but, Dixon asserts, ‘one is hard put to it to find a great general or admiral who was not conforming, submissive to authority, punitive, sexually inhibited, over-controlled, ethnocentric, anti-intellectual, assailed by doubts as to his virility, anal-obsessive, superstitious, status-hungry, rigid, possessed of a “closed” mind, and saturated in discipline’.¹² This splendidly provocative conclusion has generated considerable debate over the years, but its utility in the case of our Ambon case study is limited.¹³ With perhaps one exception, we lack the documentary evidence and oral history that might sustain claims about the personality types of key decision-makers and their propensity, if any, to authoritarianism. My analysis thus focuses on what we might learn from applying Kahneman’s analysis of decision-making, and his concepts of System-1 and System-2 thinking, to the development of planning about the defence of Ambon, as revealed in the the archival record.

Kahneman’s systems 1 and 2

It is not possible in this essay to do full justice to Kahneman’s complex book, which distils not only his lifetime of research but also the findings of multiple other psychologists, including his close colleague, Amos Tversky. For our purposes, the key arguments concern System 1 and System 2—what Kahneman calls the ‘fictitious characters’ in decision-making processes. System 1 ‘operates automatically and quickly, with little or no effort and no sense of voluntary control’.¹⁴ Continuously generating impressions, intuitions, intentions and biases, it maintains and updates a model of the individual’s personal world, which represents what is normal.¹⁵ Given the inherently confirmatory bias of System 1, people making decisions seek data that are likely to be compatible with the beliefs they currently hold. When they believe a conclusion is true, they are also ‘very likely to believe arguments that appear to support it, even when these arguments are unsound’.¹⁶ System 1, Kahneman further argues, also maintains a model of the personal world that is:

constructed by associations that link ideas of circumstances, events, actions, and outcomes that co-occur with some regularity, either at the same time or within a relatively short interval. As these links are formed and strengthened, the pattern of associated ideas comes to represent the structure of events in your life, and it determines your interpretation of the present as well as your expectations of the future.¹⁷

System 2, in contrast, ‘allocates attention to the effortful mental activities that demand it, including complex computations … [Its] operations are often associated with the subjective experience of agency, choice and concentration’. It is the conscious reasoning self that ‘has beliefs, makes choices and decides what to think about and what to do’.¹⁸

Generally, people believe that their decision-making is the domain of System 2—it is ‘who we think we are’.¹⁹ But, Kahneman argues, System 2 is slower, and this is not where ‘the action is’. At the centre of the storm is System 1, which ‘effortlessly originat[es] the impression and feelings that are the main sources of the explicit beliefs and deliberate choices of System 2’.²⁰ To be sure, System 2 has some ability to change the way System 1 works, ‘by programming the normally automatic functions of attention and memory’.²¹ Indeed, one of its main functions is ‘to monitor and control thoughts and action “suggested” by System 1, allowing some to be expressed directly in behaviour and suppressing or modifying others’.²² But System 2, which demands effort and acts of self-control is a ‘lazy controller’. Efforts of will and self-control are tiring, depleting and unpleasant.²³ Indeed, if a satisfactory answer to a hard question cannot be found quickly, System 1 slips into a mode of substitution, answering easier questions than the one asked.²⁴ Notably, in situations of uncertainty, System 1 bets on an answer, and these bets are guided by experience.²⁵

Committing Gull Force to Ambon, 1941

We can see both modes of thinking at play throughout the decision-making processes shaping the strategy of defending the islands to the northwest of Australia, but System 1 became progressively more dominant as the strategic crisis deteriorated (See Map 3.1).²⁶ The initial decision to send forces to the NEI was made in February 1941 at a Singapore conference of British, Dutch and Australian representatives (with the Americans as observers). Here, Australian authorities agreed that instead of the 8th Division of the 2nd AIF being sent to the Middle East, where the other three Australian divisions were being deployed, it would be split between Malaya, Rabaul and Darwin. While Lark Force would be sent in March 1941 to Rabaul, Gull and Sparrow Forces were to be held in Darwin in readiness to go to Ambon and Dutch Timor in the event of war breaking out with Japan. The RAAF would also contribute units to the defence of the Ambon–Timor–Darwin triangle.²⁷

Map 3.1: The Japanese offensive in Southeast Asia, December 1941 – March 1942

This decision had elements of System-2 thinking: it was a seemingly rational adjustment of planning in the face of a changing and deteriorating strategic environment. Signals intercepts in early February 1941 suggested that the Japanese might have decided to push southward, even if this meant war with the British Empire. Moreover, Britain’s capacity to provide the promised air and naval forces for the defence of Malaya and Singapore was clearly much reduced owing to the extension of the war to North Africa from late 1940 and the preoccupation of the Royal Navy with the Mediterranean and the Atlantic theatres after the fall of France. The support of the US, meanwhile, remained very ambiguous, unless it, too, were attacked by Japan. Given the geographical importance of the NEI as a barrier between Malaya and northern Australia, it was logical and rational to do everything possible to bolster the Dutch capacity to resist. The commitment to Timor and Ambon, which, like Rabaul, offered strategic airfields and harbours, was also a political gesture to the Dutch administration in the NEI. The Vichy French authorities in Indochina, after all, had succumbed to Japanese pressure.

Yet, if this was System-2 thinking, the decision to commit to providing Australian forces for Ambon and Timor was also imbued with impressions arising from System 1. The deployment of these Australian troops in single battalions, which violated the fundamental principle of concentration of force, made sense only in two circumstances: first, if the NEI were to be a second line of defence behind Malaya and British naval power at Singapore; and, second, if the Dutch military forces in the NEI had some prospect, with Australian support, of countering a Japanese onslaught if it happened that the NEI, rather than Singapore, became the front line of defence.

Both these scenarios were questionable by early 1941. The defences of Singapore were so weak that in January 1941, Prime Minister Robert Menzies went to London to lobby the British Government on the matter. As for the Dutch military capacity, it seems that Australian decision-makers had little data that might have led them to be confident about the strength of the NEI military forces in early 1941. On the contrary, since October 1939, the Dutch authorities had been desperately exploring the possibilities of securing defence supplies: modern tanks, artillery and aircraft from the UK and small arms ammunition from Australia, munitions that Australia was only partly able to provide because of its own tight supply situation.

The Australian decision to assume some responsibility for the defence of Ambon and Timor was, thus, anchored in assumptions, beliefs and impressions about the presumed superiority and permanence of imperial power. British and Dutch imperialism in the Asia-Pacific region was the normality that System-1 thinking was predisposed to maintain and update. Yes, there were gnawing doubts about the Singapore strategy, and little was known about Dutch colonial defences, but System 1’s confirmatory bias served to seek information that was compatible with the beliefs the Australian leaders held. Predisposed to discount the Japanese on racial grounds—one of Dixon’s four elements of military incompetence—Australian strategists slipped into the overconfident judgements that are typical of System-1 thinking. Horner is right to say that they were ‘locked into reliance upon the Singapore strategy’.²⁸ However, the ‘locking’ arose partly from the systemic errors of System-1 thinking, which inhibited any exploration of a radically different paradigm to imperial defence.

Sending Gull Force to Ambon

The power of System-1 thinking became increasingly evident as Australian leaders failed to review the plans to deploy to the NEI in the months that followed the original February 1941 agreement—this, despite the Allied global strategic position deteriorating markedly throughout the year. The initial British successes against the Italians in North Africa were reversed after the intervention of the German Afrika Korps in February–March 1941. Disastrous losses, including many personnel of the Australian 6th Division, were suffered in Greece and Crete in April and May, and Japan extended its control over French Indochina. While Germany’s attack on the Soviet Union on 22 June 1941 would ultimately be to the Allies’ immense advantage, the Red Army’s losses in the first months of Barbarossa were catastrophic. The British and American decision to support the Russians with munitions threatened to make resources for Singapore even scarcer. Furthermore, as 1941 progressed, British and American strategic priorities shifted away from the Pacific to the European war, with the ‘Germany First’ agreement in March 1941—whereby priority would be accorded to defeating Germany if war broke out with Japan—and the deployment of US warships to escort British convoys crossing the Atlantic.

To add to this, the feasibility of sending Australians to the NEI looked questionable on closer examination. In May and October 1941, the commander of the 23rd Brigade, ER Lind, and the commanders of Gull and Sparrow Forces conducted reconnaissance trips to Ambon and Timor. Their subsequent reports to Army Headquarters in Melbourne recommended that the forces would need Bren carriers and guns, and field or mountain artillery, since the Dutch artillery at Ambon at least seemed to be allocated to ‘several alternative roles’ (what these were was unstated). In July, Lind expressed his concerns directly to Sturdee; then, in October, he complained that the preparation by the Dutch authorities was ‘far from satisfactory’. He felt:

strongly that no part of our forces should be committed in either Timor or Ambon until the Dutch show by results that they attach as much importance to the well-being and safety of our troops as we ourselves do.

To send the forces without supporting artillery would be ‘grossly unfair to the troops’.²⁹

For all this, the decision to commit the battalion forces to support the NEI was not reviewed when war with Japan broke out. When the Japanese sank the Prince of Wales and Repulse on 10 December, any notion that the NEI could shelter behind a British naval shield at Singapore was shattered (as events would soon show—see Map 3.1). Gull and Sparrow Forces were nevertheless despatched without hesitation, their operational instructions being simply ‘to strengthen the existing defences of the operational bases at Ambon and Koepang’.³⁰ While part of No. 13 RAAF squadron of Hudson bombers proceeded to Ambon, Gull Force left Darwin on 13–14 December, three days after the sinking of the British capital ships. The Australians joined about 2,600 Royal Netherlands East Indies Army (KNIL) troops, comprising several small and understrength companies of Indonesian troops, mainly officered by Dutchmen and supported by some coastal artillery.³¹

We have some insight into the decision-making processes of the Australian chiefs of staff at this critical time from the appreciations they prepared for the War Cabinet on 11 and 15 December 1941.³² The former stated that the Japanese attack on Malaya three days earlier might well be a first step in the Japanese plan for a major attack on Australia. Moreover, this attack might take the form ‘of a direct move on Australia via the islands to the north and north-east’, including Rabaul, Port Moresby and New Caledonia. Hence, the chiefs recommended that these areas be held with a minimum of a brigade group with anti-aircraft and coastal defences. In addition, the forces needed to meet an invasion of Australia should be raised and trained (a recommendation Horner notes trenchantly that ‘would have been an admirable aim a year earlier’).³³

However, in the second appreciation, the chiefs of staff modified their view, stating that Rabaul should be left with its current garrison of one battalion group, Lark Force. They realised that a battalion-sized force would be too small to resist an invasion successfully but argued that it was essential ‘to maintain a forward air observation line as long as possible and to make the enemy fight for this line rather than abandon it at the first threat’.³⁴ The inherent contradiction in this approach—leave the force too small but expect it to resist invasion—smacked of System-1 shortcuts or heuristics. The idea that the forces on Rabaul should be withdrawn and concentrated with other forces, as System-2 thinking might have suggested, was ruled out: first, because this operation was thought to be hazardous; and, second, because ‘it must also be borne in mind the psychological effect which a voluntary withdrawal would have on the minds of the Dutch in N.E.I.’³⁵ The latter argument concerning the political importance of the island commitments had been influential in earlier decision-making and thus continued to shape thinking, even though it seems that at no stage in 1941—or for that matter, early 1942—was it ever subject to any serious System-2 type analysis or calculations substantiated by intelligence about morale in the NEI.

The refusal to evacuate Ambon

The logic that Lark Force was to remain on Rabaul, even though it had little chance of doing anything other than conducting a delaying action, would be applied also to Ambon and Timor. The problem for Sturdee and his Deputy Chief of Staff Sydney Rowell was that this logic was soon challenged by the commanding officer of Gull Force, Lieutenant Colonel Len Roach.³⁶ A First World War veteran and former militia officer, Roach was infuriated by Melbourne’s lack of response to the earlier reconnaissance reports Lind had submitted. As Roach left Darwin on 13 December 1941, he wrote in a highly critical manner to WJR Scott, a staff officer at Army Headquarters, Melbourne, which had operational responsibility for Gull Force. Stressing the urgent need for additional troops and equipment, including anti-tank guns and field guns, Roach criticised the ‘unpardonable stagnation’ of the past seven months. This pointed to ‘a policy of “wait and see” prevailing over reality and expediency’. He warned that ‘[i]f any of my excellent fellows do not arrive at their destination it will not be a case of “gallant sacrifice” but of murder due to sheer slackness and maladministration’.³⁷

Once on Ambon, Roach wrote again to Melbourne, on 17 December and on 23 December, reiterating the deficiencies in Gull Force’s arms and equipment, and the ‘imperative’ need for guns and machine guns. Thanks to a clumsy system whereby Gull Force’s lines of communications went through 7th Military District in Darwin even though the force was under direct operational control of Army Headquarters in Melbourne, some of Roach’s messages took many days to be read. Increasingly frustrated, he signalled on 24 December that the combined army forces on Ambon were inadequate to hold vital localities for more than a day or two ‘against determined attack from more than one direction simultaneously’. Even with the additional requested supplies, Gull Force’s position would be ‘precarious’ if it lacked air and naval support.³⁸ The NEI forces, Roach reported three days later, had only five days of ammunition supply and proposed to vacate five forward posts in the event of simultaneous attack from two or three directions. ‘[A]dmittedly,’ he conceded, ‘they have not very favourable alternatives with their present strength.’³⁹

On 26 December, Rowell replied: while Bren carriers would arrive on Ambon in the first week of January 1942, the additional units Roach requested were ‘not repeat not available’. The task of Gull Force then was, ‘in cooperation with local Dutch forces … to put up the best defence possible with the resources you have at your disposal’. Roach replied on 29 December that all on Ambon were ‘most eager [to] administer salutary punishment’. He felt ‘confident enemy will waver before Australian fire and bayonets’. But it seems that Roach doubted his own heroic (System-1) rhetoric. Gull Force’s intelligence officer, Captain Rod Gabriel, later recalled that when Rowell’s message of 26 December arrived, Roach placed his hand over his metal badge of rank on his shoulder and said, ‘This metal means nothing to me, but the lives of over 1100 men do’.⁴⁰ Then, on 1 January 1942, Roach wrote to Scott:

I find it difficult to overcome a feeling of disgust, and more than a little concern at the way in which we have seemingly been ‘dumped’ at this outpost position, in the first place without any instructions whatever … and in the second place with (so far) a flat refusal to consider any increase in fire power and the number of tps [troops], whilst the co-operation and assistance from the other two arms of the Service must be of very limited value indeed.

Repeating his prediction that Gull Force would be able to hold out for two or three days against a large-scale attack from several directions, Roach asked:

Is the intention to continue the policy of allowing small Forces, inadequately equipped for their task, to be spread over a vast area, so that they can be defeated in detail? I submit that it would be far preferable, if the striking power does not actually exist for offensive action whilst these outposts are held, to ‘cut the loss’ and concentrate all available Forces (including also American and Dutch) further South for a decisive engagement’.⁴¹

Roach admitted that he was speaking ‘very frankly’, but declared, ‘I would feel very guilty if I did not do so, and many valuable lives were thus lost due to the fact that your HQ had not been made fully aware of the situation’. He asked that his letter be escalated to the Minister.

Soon, Japanese air attacks on Ambon began, with seven flying boats dropping 33 bombs at three points on the island on the night of 6 January. Air raids occurred almost daily thereafter. Roach signalled on 11 January that ‘We are all completely in the dark … Prospects are gloomy’.⁴² Then, two days later, he signalled Melbourne recommending the immediate evacuation of Gull Force. The support he had requested could not be expected, and if Ambon were to be attacked by combined enemy forces such as were then operating in the Menado area, the men of Gull Force could not hold out for more than a day. It would be a ‘purposeless sacrifice of valuable manpower and arms’ to leave them on Ambon.

Roach was not alone is this assessment. Several senior officers on Ambon advised Melbourne that they supported his views: the 2/21st battalion’s second-in-command, Major Ian Macrae; the senior RAAF officer commanding Australian aircraft on the island, Wing Commander ED Scott; and the Area Combined Headquarters at Halong.⁴³ The latter sent a signal to Melbourne on 11 January pointing out that Japanese air bases were now established about 360 miles away and urgently requesting immediate reinforcement with fighters and dive bombers. Without these, only ‘token resistance [would be] possible with [the] present unsuitable aircraft’, all of which would certainly be destroyed within a day’s action by carrier-force enemy forces.⁴⁴

However, despite the intelligence from all those on the ground, Gull Force was not evacuated. Instead, Army Headquarters ordered Roach to cease his messages at once and insisted that ‘your staunch defence will have important effect … in regard [to] future Australian Dutch cooperation’. Then, on 14 January, Roach was told by Rowell to return to Melbourne:

It is apparent from messages received at Army Headquarters since your arrival, that you have not the necessary confidence in your ability to conduct a resolute defence of Ambon in co-operation with the local Dutch forces … [You] have generally given the impression that you have accepted defeat as inevitable even before being attacked.⁴⁵

Roach’s place as commander of Gull Force could be taken by Scott, who had actually volunteered on 11 January to go to Ambon and lead Gull Force.

It might be argued that these decisions were justified on the rational grounds that Gull Force’s presence at Ambon bought time for the defence of the Australian continent. Indeed, Scott said in a 1946 report:

[Gull Force] did in fact hold up a complete Japanese Division with its transports and adequate Naval and Air support for at least two weeks and inflicted heavy loss upon the enemy … [H]ad ‘Gull Force’ been withdrawn from AMBON before the Japanese attack, the Japanese Division with its Navy and Air Force support could have proceeded straight to Darwin … I am unable to say whether this force would have been destroyed on arrival at Darwin during January/February 1942, but there can be said to be at least considerable doubt on this point.⁴⁶

However, the documentary record for early 1942 provides no evidence that Army Headquarters gave any serious thought as to how long Gull Force might resist. Nor was any consideration given to the balance sheet of such a delaying action. By one later estimate, Japanese losses in occupying Ambon were only 95 killed, 185 wounded and 1 minesweeper sunk.⁴⁷ Was this worth the loss of over a thousand Australian troops and the forfeiting of a strategic opportunity to deploy these men somewhere else? It seems that no one applied to these questions the kind of ‘effortful’ calculations that would be typical of System-2 thinking.

Similarly, it seems that no thought was given in mid-January 1942 to another System-2 calculation: namely, whether evacuation from Ambon was feasible. It might well have been impossible, given the shortage of shipping and the Japanese dominance in the air. But no one asked this. Instead, Melbourne addressed a substitute question: how could they silence the man whose predictions were so dire? It was an exemplar of System 1’s substitution of an easier question for a harder one.

The dominance of System-1 heuristics—that is, the mental shortcuts that allow decisions to be made quickly without full consideration of their implications—was also manifest in the language Sturdee invoked to justify the retention of Gull Force on Ambon. We do not have a record of what he said to his Minister when seeking approval for Roach’s recall. A draft memo prepared for this purpose was apparently not sent—a good thing given its tortured argument:

The force provided, in co-operation with N.E.I. forces, is considered sufficient to retain the island against attempted occupation on a light scale. To provide sufficient forces to withstand a major attack is entirely beyond our means. Great value should accrue, however, if the enemy is denied the island except by the employment of overwhelming force.⁴⁸

However, Sturdee said, when communicating his decision to General Archibald Wavell, the recently appointed supreme commander of the American–British–Dutch–Australian Command, that Roach was dismissed because he had made it clear since his arrival at Ambon that he had ‘no spirit to conduct a resolute defence if attacked’. His latest messages indicated a ‘probable deterioration morale all troops Ambon [sic] which may be beyond power of new Australian commander to improve rapidly’.⁴⁹ Wavell responded in kind:

So far as I can judge position at Ambon not critical and in any case I am opposed to handing out important objectives to enemy without making them fight for it. Quite appreciate feelings of lonely garrison but am sure Australians will put up a stout fight whatever happens. No doubt it is wise to change commander.⁵⁰

Some weeks later, when writing his official despatch on the operations in South-East Asia, Wavell wrote:

From the standpoint of higher strategy the situation [in early 1942] bore a distinct resemblance to that when we went to the assistance of Greece a year earlier. It might then have been more prudent to let Greece go and concentrate on holding Crete and our gains in Libya. It might possibly have been more prudent here to let the N.E.I. go and to concentrate on making Burma and Australia secure. But undue prudence has never yet won battles or campaigns or wars, and from the political point of view it would have been as unthinkable to abandon our stout-hearted Dutch allies without the utmost effort to help them as it would have been a year earlier to leave the gallant Greeks unsupported. Our assistance to Greece cost us Crete and placed us in great difficulties in the Mediterranean; our attempt to hold the N.E.I. has cost us Burma and has placed India and Ceylon in danger; but that in both instances we took the right, the only, decision, I have no doubt. Just as I still believe that our expedition to Greece was by no means the forlorn hope it may have appeared, so the Allied attempt to defend the N.E.I. might well have had a more fortunate issue. The principle of engaging the enemy as closely and as far forward as possible must be maintained at all costs and will in the end bring victory.⁵¹

This was an exemplar of System-1 thinking. The Allied failure in Greece was, by any standards, a comparison that had nothing to commend it, but Wavell reverted to it when justifying the strategy adopted in South-East Asia. This was more than an inept attempt at retrospective self-justification. In Kahneman’s terms, Wavell manifested the pattern of associated ideas that represented the structure of events in his life and determined his interpretation of the present as well as his expectations of the future.

As for ‘resolute defence’, ‘stout fight’, ‘lack of ‘spirit’ and ‘lonely garrison’: these were distancing clichés that spoke to System-1 assumptions about what were appropriate qualities for military personnel to display in time of war: honour, courage, determination and a willingness to die. Australian soldiers were meant to put up the best fight possible with whatever resources could be spared. Was this not what their forebears had done at Gallipoli and on the Western Front? Sturdee, for one, had served as engineer in highly exposed sites such as Steele’s, Quinn’s and Courtney’s posts at Gallipoli at a time when it was clear that the Allies were not going to prevail in this campaign. He had won the Distinguished Service Order for his work in 1915 and 1916. Wavell, for his part, lost an eye and won the Military Cross in the Second Battle of Ypres in 1915.

Moreover, the codes of honour invoked by Sturdee and Wavell served, as the high diction of war commonly does, to ensure that a threatening situation was ‘met by fight rather than flight’. As Dixon says, these codes do this:

by making the social consequences of flight rather more unpleasant than the physical consequences of fight. Whereas the latter might lead to physical pain, mutilation and death, the former eventuates with greater certainty in personal guilt and public shame.⁵²

The Australian authorities certainly shamed Roach. No officer at Army HQ met him on his return to Melbourne to ask what intelligence he might have about the Japanese. On 9 March 1942 he was discharged from the 2nd AIF and placed on the reserve of officers. In November 1943, Roach sought to have the operational order recalling him from Ambon on 14 January 1942 withdrawn, on the grounds that it was ‘unjustifiable and casts an unwarranted reflection on my personal attributes and my methods of command’. The Adjutant-General, Major General CEM Lloyd, did Roach the courtesy of meeting him in early December 1943, but soon wrote that the Commander-in-Chief (General Thomas Blamey) considered it impossible to withdraw an order made by an authority that shortly after ceased to exist (presumably a reference to the dissolution of American–British–Dutch–Australian Command in February 1942). Nothing in the files, Roach was told, ‘basically affects your military reputation’.⁵³

The mystery of Scott

We are left with the mystery of why Scott volunteered to assume command of Gull Force on 11 January, a time when the Japanese had already occupied Kuala Lumpur and were only 320 kilometres from Singapore. It is a decision that seems to defy rational explanation.

Scott’s earlier career was intriguing and complex.⁵⁴ He had a distinguished record of service in the First World War, serving in Gallipoli and France. Twice mentioned in dispatches, he was awarded the Distinguished Service Order for his work at Flers on 14 November 1916. Across the war years, he also suffered considerable injuries and ill health: malaria at Gallipoli in 1915; shell shock at Pozières in 1916; concussion when he fell from a horse in France in 1917; and debility in 1918.⁵⁵

After the war ended, Scott was made responsible for arranging the shipping for the repatriation of the AIF from Britain in 1919, a role that he sees to have performed with distinction. On his return to Australia, Scott became an insurance broker but he continued to be active in military affairs, joining the militia, the loyalist and anti-communist King and Empire Alliance, and the clandestine counter-revolutionary group the Old Guard. Commonly Scott is thought to have been the model for Jack Callcott, a leading character in DH Lawrence’s 1922 novel, Kangaroo. To add to the complexity, in the 1930s Scott also became an enthusiast for all things Japanese. He visited Japan in 1934, collaborated with Japanese consular officials and wrote articles for the Sydney Morning Herald praising Japanese industry and defending that country’s foreign policy. His membership of the Australia–Japan Society, an organisation suspected by the Commonwealth Investigation Branch of espionage activity, proved a complication when in April 1935 Scott joined military intelligence: he was refused access to secret files.

After the outbreak of the Second World War, Scott was appointed in June 1940 to the General Staff in Melbourne. From February to May 1941, he commanded the guerrilla warfare training centre for independent companies at Wilsons Promontory. On his return to Army HQ, he assumed the role of liaison officer for Gull and Sparrow Forces.

Scott’s offer to take over command from Roach was made in a personal and confidential handwritten note to the Director of Military Operations and Plans, which was attached as something of a postscript to an analysis of the aircraft available in the area of Gull Force. Writing at 2315 hours—perhaps not the optimal hour for System-2 thinking—Scott pondered the possible repercussions should Gull Force be the first objective chosen by the enemy:

A stubborn resistance & a good fight even against overwhelming odds now, must stiffen resistance everywhere and clinch our association with the N.E.I. Not to mention the effect on the U.S.A. Withdrawal or a weak resistance will set the pace for future threats or worse, in the near Pacific Ocean area & Australia’s itself. It could affect the action of A.I.F. troops in Malaya … I am convinced that there is a political significance behind this which may well be worth consideration at least.

Hoping that his views would be taken ‘as a very ernest [sic] opinion that much rests on this position in Ambon and not as an impertinence from a junior officer’, he claimed that he would be ‘proud indeed’ to take over command of Gull Force immediately:

I have no particular belief in my ability but I have a definite belief in my ability to inspire confidence in men and to lead them. I apologise for blowing my own trumpet in this way, but for the first time in my life I feel it perhaps justified.⁵⁶

Unlike Wavell and Sturdee’s decision-making, Scott’s carried a high personal risk, and we can only assume that he was persuaded of the political and strategic value of the Ambon operation. He had made a note in the margins of an earlier message from Roach that predicted that Gull Force could not last for more than a day or two. ‘Well worth it—what’s the force there for?’⁵⁷ But, surely, Scott’s judgement was also infused with the emotions and assumptions of System-1 thinking. He read Roach’s protests not as rational—and as it happened, accurate—intelligence about the flaws of the island strategy but as the utterances of a man unfit to hold a position of command. Scott was, after all, a political soldier. In May 1921, he had stormed a socialist gathering at Sydney’s Domain and four years later had taken instructions from Prime Minister Stanley Bruce to raise a force of 500 AIF veterans to assist police should plans to deport union leaders Tom Walsh and Jacob Johnson lead to union violence. As a senior member of the Old Guard, Scott also threatened to overthrow the New South Wales Labor government of Jack Lang. For Scott, then, the ‘normal’ world was one in which he thought politically.

There is every reason, too, to see Scott as identifying himself as a man of action. Did he find it demeaning to sit out the war in an office at headquarters? Did he think this was his last opportunity—he was already 53—to lead men in battle? His brother, Lieutenant Colonel (Alan) Humphrey Scott, had died while acting in command of the 56th Battalion at Polygon Wood in 1917.⁵⁸ Charles Bean concluded in his Official History of the AIF in 1917 (published in 1933) that had Humphrey survived, he would ‘almost certainly have risen to brigade command’.⁵⁹ Perhaps Scott hankered to emulate his lost brother’s success in command.

Possibly, too, given his experience with independent companies, Scott thought that Gull Force might retreat in the face of the Japanese attack and hold out in the more remote parts of Ambon (as the remnants of Sparrow Force and the 2/2 Independent Company did in Portuguese Timor). In 1954, Scott told the Official Historian, Lionel Wigmore, when asked to comment on a draft chapter on the Ambon campaign: ‘You must remember that I was very well aware that I was taking on a job which held out no hope of survival.’⁶⁰ Whatever his motivation, it seems clear that Scott fell victim to System-1 heuristics and failed to take the time to process all the information available to him, and to weigh it up properly.

Scott also manifested the overconfidence typical of System-1 thinking. Kahneman warns that ‘your intuitions will deliver predictions that are far too extreme and you will be inclined to put too much faith in them’.⁶¹ When Scott arrived on Ambon on 16 January, he claimed that he was ‘entirely satisfied’ with the morale of Gull Force, and ‘impressed’ with Colonel Joseph Kapitz commanding the Dutch forces.⁶² But the situation soon proved to be as dire as Roach predicted. As Japanese air raids increased from mid-January onwards, Allied aircraft were either destroyed or withdrawn. When, on 28 January, the few remaining serviceable Hudsons of No. 13 RAAF squadron were evacuated, Gull Force had no air support. On the night of 30 January, the Japanese attacked. Outnumbered and outgunned, the Dutch forces dissolved, the purported psychological value of the Australian presence proving, it seems, to have been of no use. Most of the Australians retreated, with some energetic defending actions, including on Mount Nona, to Eri, on the tip of the southern peninsula of the crab-shaped island. Three days later, with his troops lacking food, water and ammunition, and exposed to Japanese bombing and shelling, Scott surrendered. Meanwhile, across the bay, the detachment of Australian troops defending the Laha airfield were overwhelmed by a concentrated attack of enemy troops, dive bombers, fighter planes and artillery. Most were soon executed by the Japanese.

Scott then faced three-and-a-half years of humiliating captivity, a period in which he suffered a mental collapse, which might be attributed in part to the unresolved trauma of shellshock in the First World War. With considerable malice, he sought to allocate blame for his failure to others: first, the Dutch troops and Kapitz, of whom Scott wrote in his 1946 report:

It is unfortunate that there is nothing to be said which could be regarded as bringing credit either to the one or the other. My unhappy experience with these people both at Ambon and Hainan leave [sic] me convinced that there is no one characteristic which an Australian or a Dutchman could find in common.⁶³

Scott’s second target was Gull Force, the men he had thought he could lead but whom he now viewed as an undisciplined rabble. Faced with what he saw as poor behaviour within the POW camp on Ambon, Scott manifested behaviour that had some of the elements of Dixon’s authoritarian personality. In March 1942, he threatened to hand over to the Japanese for punishment Australians who were guilty of stealing, refusing to obey orders or insolence to their officers. When later interned on Hainan, he carried out this threat—with the result that the men in the camp ultimately mutinied and took discipline into their own hands, while the Australian officers debated whether to declare Scott insane and assume responsibility for management of the camp themselves.⁶⁴ When the survivors of Ambon returned to Australia, it was not Scott but Roach they chose to lead them on Anzac Day marches.

Conclusion

Any attempt to analyse the cognitive processes behind military decision-making, as this essay has done, is necessarily speculative and subjective. For one thing, the documentary evidence tracing the processes by which Australian military leaders made decisions in 1941–42 is thin. Most of Australia’s leaders in the Second World War, alas, did not keep self-reflective diaries as did the British Chief of the Imperial General Staff, Field Marshal Sir Alan Brooke. Moreover, no leader of the 1940s was ever subjected to psychological examination or interviewed by oral historians. We can only infer what might have been the assumptions and biases infusing their thinking. Nonetheless, perilous though it might be, the exploration of decision-making through the lens of later behavioural psychological theories of modes of cognition can only enrich the empiricism of military history. Understanding System-1 and System-2 thinking does not, of course, exonerate Australia’s military leaders. Horner’s judgement that Army leaders should be indicted for the island strategy still stands. But we can perhaps better appreciate how it was that these men persisted with a strategy that became increasingly irrational as the strategic situation changed in 1941–42 and that would, they ultimately knew, condemn thousands of Australians to death or captivity. Moreover, we can address the need, flagged by Horner, for military historians to acknowledge more fully the human element that shapes military decision-making in times of crisis.

---

1. 1 David Horner, Crisis of Command: Australian Generalship and the Japanese Threat, 1941–1943 (Canberra: Australian National University Press, 1978).

2. 2 David Horner, ‘Lieutenant-General Sir Vernon Sturdee: The Chief of the General Staff as Commander’, in The Commanders: Australian Military Leadership in the Twentieth Century, edited by David M Horner (Sydney: George Allen & Unwin, 1984), 149.

3. 3 David Horner, Defence Supremo: Sir Frederick Shedden and the Making of Australian Defence Policy (Sydney: Allen & Unwin, 2000), 134.

4. 4 Horner, ‘Lieutenant-General Sir Vernon Sturdee’, 154.

5. 5 Horner, ‘Lieutenant-General Sir Vernon Sturdee’, 151.

6. 6 Horner, Crisis of Command, 280.

7. 7 Horner, ‘Lieutenant-General Sir Vernon Sturdee’, 154.

8. 8 Horner, Crisis of Command, 36.

9. 9 Horner, Crisis of Command, xviii.

10. 10 Daniel Kahneman, Thinking, Fast and Slow (New York: Farrar, Straus and Giroux, 2011).

11. 11 Norman Dixon, On the Psychology of Military Incompetence (London: Jonathan Cape, 1976). For a useful review, see William Eckhardt, ‘A Review Essay’, Peace Research: The Canadian Journal of Peace and Conflict Studies 21, no. 1 (1989): 47–51.

12. 12 Dixon, On the Psychology, 347–48.

13. 13 In one critique of Dixon, Amos Perlmutter notes that since 1914, strategy is ‘fusionist, both a political and military skill’. Military incompetence is, thus, ‘no longer the sole property of generals but results from the combined efforts of inept strategists, in and out of uniform’. Amos Perlmutter, ‘Military Incompetence and Failure: A Historical, Comparative and Analytical Evaluation’, Journal of Strategic Studies 1, no. 2 (1978): 121–38, especially 121.

14. 14 Kahneman, Thinking, Fast and Slow, 20.

15. 15 Kahneman, Thinking, Fast and Slow, 71.

16. 16 Kahneman, Thinking, Fast and Slow, 45.

17. 17 Kahneman, Thinking, Fast and Slow, 71.

18. 18 Kahneman, Thinking, Fast and Slow, 21.

19. 19 Kahneman, Thinking, Fast and Slow, 415.

20. 20 Kahneman, Thinking, Fast and Slow, 21.

21. 21 Kahneman, Thinking, Fast and Slow, 23.

22. 22 Kahneman, Thinking, Fast and Slow, 44.

23. 23 Kahneman, Thinking, Fast and Slow, 42.

24. 24 Kahneman, Thinking, Fast and Slow, 25–26.

25. 25 Kahneman, Thinking, Fast and Slow, 80–81.

26. 26 The following analysis draws on my earlier research on Gull Force: Joan Beaumont, Gull Force: Survival and Leadership in Captivity 1941–1945 (Sydney: George Allen & Unwin, 1988).

27. 27 For details of negotiations between Australia and the Dutch, see Herman Bussemaker, ‘Australian–Dutch Defence Cooperation, 1940–1941’, Journal of the Australian War Memorial 29 (1996), www.awm.gov.au/articles/journal/j29/herman.

28. 28 Horner, Crisis of Command, 36.

29. 29 Brigadier EF Lind, ‘Reconnaissance Reports, Timor and Ambon, 6–12 Oct 41’, Australian War Memorial (AWM) 54, 573/6/4.

30. 30 ‘AHQ Operation Instruction no. 15, 6 December 1941’, AWM 54, 573/6/3.

31. 31 Lionel Wigmore, The Japanese Thrust, Australia in the War of 1939–1945, vol. IV, series 1—Army (Canberra: Australian War Memorial, 1957), 421.

32. 32 For details see David Horner, High Command: Australia and Allied Strategy, 1939–1945 (Canberra: Australian War Memorial and Sydney: George Allen & Unwin, 1982), 142–43.

33. 33 Horner, High Command, 142.

34. 34 Horner, High Command, 143.

35. 35 ‘Chiefs of Staff Appreciation—Defence of Australia and Adjacent Areas’, 15 December 1941, National Archives of Australia (NAA): A2671 418/1941.

36. 36 The account that follows, unless otherwise stated, draws on Wigmore, The Japanese Thrust, 420–24.

37. 37 Roach to Scott, 13 December 1941, AWM 54 573/6/3.

38. 38 Gull Force to Army HQ, 24 December 1941, AWM 54 573/6/3.

39. 39 Roach to AHQ Melbourne, Disposition Gull and NEI Forces, 27 December 1941, AWM PR 8/3/84.

40. 40 Comment concerning Col. Roach provided by Gabriel to Joan Beaumont, February 1986.

41. 41 Roach to Scott, 1 January 1942, AWM 54 573/6/10.

42. 42 ACH Halong to CWR Melbourne, A.12/11/1, 11 January 1942, AWM 54 573/6/10.

43. 43 Macrae to Scott, 1 January 1942, ACH Halong to CWR Melbourne, 13 January 1942, AWM 54 573/6/10.

44. 44 Wigmore, The Japanese Thrust, 423.

45. 45 Rowell to Roach, 14 January 1942, AWM 54 573/6/10.

46. 46 Lt-Col. WJR Scott, ‘Introduction’, in Report on Ambon and Hainan, 8 May 1946, AWM 54 573/6/1A.

47. 47 National Defense College of Japan, The Operations of the Navy in the Dutch East Indies and the Bay of Bengal, trans. William Remmelink (Leiden: Leiden University Press, 2018), cited in ‘Battle of Ambon’, Wikipedia, en.wikipedia.org/wiki/Battle_of_Ambon.

48. 48 Draft memo, Chief of the General Staff to Minister, January 1942, AWM 54 573/6/10.

49. 49 Sturdee to Wavell, 14 Jan. 1942, AWM 54 573/6/10. This was repeated to the Dutch commander of the KNIL, General Hein ter Poorten.

50. 50 Wigmore, The Japanese Thrust, 424.

51. 51 General Wavell’s Despatch on Operations in South-West Pacific, 15 January – 25 February 1942, WO 10/2556, National Archives, London.

52. 52 Dixon, On the Psychology, 197. I owe the term ‘high diction’ to Paul Fussell, The Great War and Modern Memory (Oxford: Oxford University Press, 1977), esp. 21–22.

53. 53 This correspondence is found in AWM 54 573/6/10.

54. 54 See Andrew Moore, The Secret Army and the Premier: Conservative Paramilitary Organisations in New South Wales 1930–32 (Sydney: University of NSW Press, 1989) passim; Moore, ‘Life Summary: William John Scott’, Australian Dictionary of Biography (Canberra: The Australian National University, online edition), adb.anu.edu.au/lifesummary/scott-william-john-8373.

55. 55 ‘Scott, WJR’, B2455, NAA.

56. 56 Scott to DMO&P, 11 January 1942, AWM 54 573/6/10.

57. 57 Annotation on Roach telegram, Gull “G” & “I” matters, 24 December 1941, AWM 54 573/6/3.

58. 58 ‘Scott Alan Humphrey’, B2455, NAA.

59. 59 CEW Bean, The A.I.F. in France, 1917, vol. IV of The Official History of Australia in the War of 1914–1918, (Sydney: Angus & Robertson, 1933), 836n.

60. 60 Scott to Wigmore, 15 June 1954, AWM 67 3/353.

61. 61 Kahneman, Thinking, Fast and Slow, 194.

62. 62 Cited in ‘Wavell from Sturdee’, 18 January 1942, AWM 54 573/6/10.

63. 63 Scott, ‘Introduction’.

64. 64 See Beaumont, Gull Force, 96–97, 180–85.