한자로

Rochlin: Trapped in the Net -- Chapter 9: Unfriendly Fire

The Wayback Machine - https://web.archive.org/web/20060905022747/http://www.pupress.princeton.edu/books/rochlin/chapter_09.html

All rights reserved. No part of this book may be reproduced in any form by any electronic or mechanical means (including photocopying, recording, or information storage and retrieval) without permission in writing from the publisher, except for reading and browsing via the World-Wide Web.

Chapter 9: Unfriendly Fire

[¶1.]

Introduction

[¶2.] In 1868, the U.S. Navy launched the Wampanoag , the most advanced naval weapon of her time. ¹ Steam-powered, propeller-driven, with a hull designed primarily for speed, she was heavily armed with ten 8-inch rifled guns, a 60-pounder, two 100-pounders, and four howitzers. During her trials she maintained speeds of over sixteen knots for over twenty-four hours in seas that were at times heavy. There was no warship afloat that could match her, and there would not be for nearly twenty years thereafter.

[¶3.] The Wampanoag demonstrated her seaworthiness and efficiency in a year of service, during which time her officers and crew could find no major faults at all. Nevertheless, a naval board inspecting her a year after her commissioning declared on the basis of their opinions, and in the face of the evidence, that she was a problematic and unseaworthy ship, "a sad and signal failure, and utterly unfit to be retained in the service." ² However outrageous they now seem, the board's claims and findings were never successfully challenged. After being laid up for a year, she was transferred to noncombat duty for several years before ultimately being decommissioned and sold. It was not until the 1880s that the U.S. Navy was to move beyond sail to the first steam-powered and steel-hulled warships.

[¶4.] The story of the Wampanoag is often told as an example of bureaucratic resistance to change in the military. Traditional militaries were islands of stability in time of peace. A great premium was placed on compliance and routine, on keeping a low profile: "The Navy officer was an organization man. He spent his life both obeying and giving orders within an institutional context, moving up gradually through the ranks, preserving and identifying with the status quo, honoring tradition, defending the organization that provided him with security and recognition." ³

[¶5.] The usual argument is that what the Navy sought was to maintain stability, to arrest change so rapid that it would dislocate both the military bureaucracy and the civil economy with which it interacted. ⁴ In this frame, the case of the Wampanoag is not substantively different from historical cases of resistance to naval radio, submarines, precision gunnery, and automatic weapons around the turn of the century, or to the tank, the submarine, and the bomber between the two world wars. ⁵

[¶6.] On second look, however, the Navy's decision on the Wampanoag is more complex than simple resistance to change. According to Elting Morison, the concluding section of the Hearing Board's report reflects a real concern, and not just bureaucratic inertia or a desire to preserve tradition:

[¶7.]

Lounging through the watches of a steamer, or acting as firemen and coal heavers, will not produce in a seaman that combination of boldness, strength, and skill which characterized the American sailor of an earlier day; and the habitual exercise by an officer of a command, the execution of which is not under his own eye, is a poor substitute for the school of observation, promptness, and command found only on the deck of a sailing vessel. ⁶

[¶8.] In the 1870s, at the threshold of the most rapid change in naval technology in history, the board feared the possibility that radically new weapons and systems would seriously disrupt patterns of social and organizational behavior in ways they could neither foresee nor control. Morison goes on to remark:

[¶9.]

What these officers were saying was that the Wampanoag was a destructive energy in their society. Setting the extraordinary force of her engines against the weight of their way of life, they had a sudden insight into the nature of machinery. They perceived that a machine, any machine, if left to itself, tends to establish its own conditions, to create its own environment and draw men into it. Since a machine, any machine, is designed to do only part of what a whole man can do, it tends to wear down those parts of a man that are not included in the design. ⁷

[¶10.] In his repetition of the story, Deitchman points out that similar concerns persist in the modern age, where the power, complexity, extent, and rate of technical change are immeasurably greater than they were in the 1860s, or the 1930s. Although he makes the familiar argument that the current situation is exactly opposite to that surrounding the case of the Wampanoag , that our defense "now depends on high-technology industry that is built on the concept of continual change," ⁸ he also notes that the systemic, technological changes that will ensue will not come easily or quickly, or without pain or the possibility of fundamental surprise.

[¶11.] Militaries share with civilian organizations that manage safety-critical technical systems a deep concern over the possibility of unanticipated side effects or unexpected consequences, particularly when they affect the performance of their troops. The old slogan "we train like we fight" expresses perfectly the purpose of military training--to make behavior reliable and reasonably predictable in combat.

[¶12.] Because the new, far more complex and sophisticated equipment now being put into place is much more demanding of the time, attention, and training of the people who man it, it is even more likely to create its own environment and draw its operators into it. The next two chapters will discuss the historical evolution both of automated and computerized military equipment and of the use of new information and communication techniques for advanced command and control. This chapter provides context by setting out in brief three examples where the consequence was not only surprise but tragedy. In all three cases, the electronic warfare systems designed to enhance warnings and facilitate engagement actually contributed to the errors in cognitive framing that occurred, both directly through changing the nature and balance of the process of decision and indirectly through the changes that electronic warfare has made in the combat environment.

[¶13.]

A "Reasonable Choice of Disaster" ⁹

[¶14.] On February 21, 1973, two jet fighters of the Israeli Air Force intercepted a Libyan airliner on its way to Cairo. For a number of reasons, the plane did not realize that it had wandered off track, on a day when a sandstorm over Egypt and the Sinai blocked direct visual verification of course. Instead, it was heading over disputed territory in the general direction of Beer Sheva and Redifim--an Israeli air base near Bir Gafgafa.

[¶15.] Unaware that they had wandered off track, and insensitive to the potential dangers of flying over neighboring Israeli territory, the Libyan crew consistently misread almost every signal they were given by Cairo or the Israeli fighters right up to the moment when they attempted their emergency landing. They did not realize they were lost, even when Cairo airport informed them that it had no idea where they were. They misidentified the Israeli Phantom jets as Egyptian Mirages; even when warning shots were fired, their response was to call Cairo and complain. They took none of the standard actions agreed upon for hostile intercepts (such as lowering the landing gear).

[¶16.] When the Libyan plane failed to respond in return, the Israelis flew close by, reporting that "all the window shades were down"--in itself an unusual and somewhat suspicious appearance. After the first warning shots, the Libyan crew did lower the gear and descend toward Redifim. Mistaking it for Cairo East, an Egyptian military base, they climbed out again seeking to fly westward to the more "correct" international airport at Cairo West.

[¶17.] Once the airliner re-ascended from its approach to Redifim, the Israelis concluded that it was trying to escape. The air force commander decided that it was a hostile flight, and that allowing it to escape would encourage another attempt. He therefore ordered the fighters to shoot at the base of the wings to force it down immediately. ¹⁰ The Israelis fired again, but the liner continued to fly westward. At this point, the Israelis fired once more, hitting the base of the wing. The airliner tried to land on the flat sand but failed. After a short run on the ground it crashed, killing all but six of the 116 people aboard.

[¶18.] The Libyan crew seemed somewhat confused (the captain and engineer were French, the copilot a Libyan who spoke little of it). The behavior of the Israelis was composed and consistent, even if it did not correspond to the "realities" that emerged after the cockpit tapes of the doomed airliner were reviewed. The Israeli pilots did identify the airliner as Libyan, which meant potentially hostile. Their aircraft were clearly marked with the Star of David, and did execute the maneuvers internationally agreed upon--signaling and rocking their wings.

[¶19.] Lanir points out that the sequence of events was highly improbable statistically, making the event something that could not have been anticipated or foreseen--as much of an "inconceivable occurrence" as having an Iranian airliner pass over the USS Vincennes in the middle of a surface fire fight, or a Korean airliner turn up over Kamchatka. Both sides acted rationally and correctly within the (erroneous) cognitive frames in which they were made.

[¶20.] Any cognitive framing is based on an initial set of assumptions and distinctions. Once the frame is set, these are difficult to challenge, particularly under stress, and most particularly when accompanied by pressure to act quickly. Lanir argues that the cognitive frame of the Israelis was largely predetermined by their own electronic command and control system. The fighters were not sent into the air to explore an uncertain and presumably unknown situation, but to carry out a mission.

[¶21.] The Israeli Defense Force is proud of having highly trained and independent fighting units capable of acting flexibly without central control. But the presumed discretion of the pilots to act on their own judgment and initiative turned out to be illusory. Instead, their actions and responses were framed by the causal logic of the early warning system. The unanticipated outcome of the electronically integrated system was to create a disaster out of otherwise "reasonable choices," rather than from the human errors the system was put in place to avoid. ¹¹

[¶22.]

The USS Stark

[¶23.] The attack by an Iraqi Mirage jet on the USS Stark during the Iran-Iraq war is in many ways similar. On May 17, 1987, the Stark was on routine patrol in the Persian Gulf to protect neutral shipping. At about 8:00 a.m., a long-range U.S. electronic warning and control aircraft (AWACS) picked up an F-1 Mirage, positively identified it as an Iraqi aircraft, and passed the notification on to U.S. Naval units operating in the Gulf. ¹² A little after 9:00 that morning, the aircraft was picked up as an unknown on the Stark 's radar, at a range of about seventy miles.

[¶24.] Under the Rules of Engagement then in force, the Stark was not compelled to wait out an attack. Three days earlier, an Iraqi aircraft had come within forty miles of the USS Coontz . The Coontz not only transmitted an interrogation, it turned broadside to the aircraft, unmasked its radars, armed its chaff dispensers, and mounted an anti-aircraft missile. The Mirage eventually closed to ten miles, but never turned toward the ship (it went on to attack a tanker).

[¶25.] Once the Mirage had closed to less than seventy miles, the Tactical Operations Officer (TAO) of the Stark was tracking it continuously. When the aircraft closed to thirteen miles, the Stark identified itself by radio, and requested identification from the aircraft. A second inquiry at a range of eleven miles also brought no response. At about 9:11, the operator of electronic intercept equipment aboard the Stark reported that it had been locked onto by the aircraft's fire control radar.

[¶26.] When the TAO discovered the lock-on by the Mirage's radar, he immediately started to bring the ship's Phalanx close-in gun system up. He also requested a lock by the ship's air defense radar. However, the attack was coming in over the port bow, and the primary radar was blocked by the superstructure. At 9:12, the TAO ordered a secondary radar brought up, but before it could be activated an Exocet missile hit the ship, severely damaging it and killing thirty-seven aboard. A second missile impacted shortly thereafter. It ignited a large fire in the aluminum superstructure that was not put out for some time. The ship had neither taken evasive maneuvers nor brought its defensive weapons systems to bear.

[¶27.] Subsequent to the U.S. Navy's own inquiry, the Staff Report of the Committee on Armed Services concluded that although the Rules of Engagement allowed for a more aggressive defensive posture, the real world was more difficult. ¹³ At that time, Iraq was considered a near-ally against Iran, and had never attacked a U.S. ship despite several opportunities.

[¶28.] In all probability, the incident had been caused by complementary errors of interpretation. The Iraqi attack was probably inadvertent, given that their cognitive frame was one of expected hostility. In the era of electronic warfare, the fear that he who hesitates is almost certainly lost leads to a policy of attacking immediately almost anything the radar engages. In contrast, the Stark suffered from being caught in a cognitive frame in which an attack was unexpected. They regarded the closing of the Mirage as a puzzle rather than a threat, and did not take action to unmask its defensive systems in time for them to engage. ¹⁴

[¶29.] Was this, too, an avoidable incident from which useful lessons could be learned? Perhaps. However, the lessons that were thought to have been learned by the U.S. Navy were not about cognitive disjunctures, but about quick response. This was to be a contributing factor in framing an even more consequential event in the Gulf the following year.

[¶30.]

Tragedy over the Persian Gulf

[¶31.] In July 1988, the USS Vincennes was patrolling the restricted waters of the Persian Gulf, as she had been for several weeks as part of the fleet enforcing the U.S. embargo of Iran. ¹⁵ Also in the vicinity were two U.S. frigates, the USS Elmer Montgomery (FF 1082) and the USS Sides (FFG 14). Like many of the U.S. ships involved in the Gulf patrol, all three were designed and built to be blue-water fighters, part of the U.S. global maritime strategy. Bottled up in the Straits of Hormuz, these billion-dollar bundles of sophisticated and advanced technology were mostly worried about low-technology attacks from mines and Iranian speedboats.

[¶32.] Like other Ticonderoga -class Aegis cruisers, the Vincennes is a fast, lightly armored ship built on a long, narrow destroyer hull. Although armed with various surface-to-surface guns and a variety of systems for close-in air defense, she was optimized for providing blue-water air defense to an aircraft carrier battle group. Her real "main battery" of standard SM-2 anti-aircraft missiles stored deep in her magazines are controlled by the advanced Aegis electronic fire-control system around which she was built.

[¶33.] Aegis was designed to be capable of processing, interpreting, and displaying the data from a complex air battle extending over many hundred square miles, and displaying it on an enormous visual display in the Command Information Center (CIC). In order to perform effectively, the electronic suite must be able to track and distinguish friendly and potentially hostile aircraft at ranges of tens of miles while engaging a variety of potential targets ranging from high-flying reconnaissance aircraft to high-speed cruise missiles. The resulting array of computers, modern displays, and other data gathering and command posts being too large for the narrow hull, it is located up in the superstructure behind the phalanx of huge phased-array radars.

[¶34.]

Surface Attack

[¶35.] With the memory of the attack on the USS Stark the previous year still fresh in every sailor's mind, all weapons and warning systems on the Vincennes were up and fully manned. On July 2, several armed small boats of the Iranian Revolutionary Guard Corps (IRGC) had positioned themselves at the western approach to the Straits, and were challenging merchant vessels. Late that day, the Montgomery had come close enough to a ship attack in progress to fire warning shots at several of the IRGC boats.

[¶36.] Earlier on the morning of July 3, the Montgomery , at the northern end of the Straits, reported another attack by seven small IRGC boats armed with machine guns and rockets. Shortly thereafter, another wave of thirteen such boats was reported, in three groups, one of which took a position of the Montgomery 's port quarter. At 7:42 a.m. local time, the Vincennes was dispatched to the area to investigate the situation. At about 9:45 a.m., one of her helicopters sent out to monitor the situation having been fired upon, the Vincennes went to General Quarters and took tactical command of the Montgomery .

[¶37.] As they approached the position of the boats, several were observed to turn toward the U.S. ships and close in a threatening manner. Taking this as prima facie evidence of hostile intent, Middle East Joint Task Force Command gave permission to engage, and the Vincennes opened fire at 10:13 a.m., starting a surface meleé that was to continue throughout the incident. The IRGC boats, fully aware of the advantage conferred by their small size and maneuverability, did not flee, but turned to engage, hoping thereby to inflict some damage on the far more expensive and valuable U.S. ships.

[¶38.] Four minutes after she opened fire, a bad round fouled the Vincennes' s forward 5" mount. Because she was armed with only one gun fore and one aft, the TAO was forced to maneuver the ship radically--using 30 degrees of rudder at a ship's speed of 30 knots--to bring the after 5" mount to bear on the most threatening of the small boats. This caused the ship to heel dramatically, sending loose equipment flying. Because the CIC, which contains not only the Aegis displays but all other displays and consoles from which the ship is fought, is high above the water line, the effect was particularly dramatic. Books, publications, and loose equipment went flying off desks. Desk and file drawers flew open. Many of those on duty had to grab for the nearest support to avoid being thrown to the deck.

[¶39.] The surface engagement ended inconclusively at 10:33, the Vincennes having expended seventy-two rounds of 5" ammunition and the Montgomery 47. The IRGC boats were then in retreat, one of them having been sunk by U.S. gunfire. But in this relatively short period of time, the Vincennes had detected, identified, and shot down an Iranian airliner that had the bad luck, or bad judgement, to overfly the scene of battle.

[¶40.]

Iran Air Flight 655

[¶41.] That same morning, the captain and crew of Iran Air Flight 655 were at Bandar Abbas airfield in southern Iran, preparing for a routine 150-mile flight over the Gulf to Abu Dhabi. It was one of the many ironies of the half-war then being fought in the Gulf that such commerce proceeded almost routinely, in, around, and through what amounted to open combat zones. Even the status of Bandar Abbas itself was ambiguous. Several Iranian F-14 fighters had arrived only a day or two earlier, and their purpose and mission on what was primarily a civil airfield was never made entirely clear.

[¶42.] Like most modern aircraft, the Iranian airliner was equipped with an aircraft identification transponder, a modern form of the old "identification, friend or foe" (IFF) system of World War II. When interrogated by a radar signal from a potential adversary, the transponder "squawks" (gives off a specific response signal) in a prespecified, fixed mode. The Iranian F-14s at Bandar Abbas are presumed to have been set to squawk in "Mode II," a mode that would identify to the U.S. ships that the aircraft in question were military, and Iranian. Being a commercial flight, Iran Air 655 was instructed to squawk in Mode III, a signal that identifies civilian traffic. A unique transmission code number, 6760 in this case, was assigned to distinguish this particular flight from others.

[¶43.] Scheduled departure time was 9:59 a.m. local time. The usual mixed load of businesspeople and relatives was making the quick hop to the comparative peace and luxury of the eastern Gulf. The flight was assigned routinely to commercial air corridor Amber 59, a twenty-mile-wide lane on a direct line to Dhubai airport. Owing to the short distance, the flight pattern would be a simple trajectory--climbing out to an altitude of 14,000 feet, cruising for a short time, then descending gradually into Dhubai.

[¶44.] The matter of life or death for those aboard seems to have been settled by a still unexplained (but not uncommon) eighteen-minute delay in departure. After taking off from runway 21, Flight 655 was directed by the Bandar Abbas tower to turn on its transponder and proceed over the Gulf. Because of the delay in takeoff, it appeared on the Vincennes 's radar at 10:17, just after the duel with the IRGC patrol boats had begun. At 10:19, the Vincennes began to issue warnings on the Military Air Distress frequency, and at 10:20 began warnings on the International (civil) Air Distress frequency as well. It was just at this moment that the TAO ordered the sharp turn that created havoc aboard.

[¶45.] During the next three minutes, with the ship in a radical heel, the CIC in confusion and disorder, and while continuing to engage the IRGC boats, the Vincennes issued a number of warnings on both military and civil distress frequencies, it (mistakenly) identified the Airbus 320 as a possible Iranian F-14, it (mistakenly) reported hearing IFF squawks in Mode II, and it (mistakenly) reported the aircraft as descending toward the ship when it was in fact still climbing according to its usual flight plan. ¹⁶

[¶46.] Having informed Joint Task Force Command that a potentially hostile aircraft was rapidly closing to within potential missile attack range, the Vincennes received permission to engage. Captain Rogers, the Commanding Officer (CO), held out for a minute or two more, by which time the still unidentified aircraft had closed to fifteen miles and was still being reported as descending toward his ship. At about 10:24 a.m., seven minutes into the Iranian Airbus's flight, and eight minutes into Vincennes 's fire fight, the CO fired two SM-2 standard missiles at the unknown target.

[¶47.] A few seconds later, with the Airbus still on its assigned climbout, and slightly to one side of, but well within air corridor Amber 59, it was intercepted by one or both of the missiles at a range of eight nautical miles and an altitude of 13,500 feet. Flight 655, with some 290 people from six nations aboard, tumbled in flames into the Persian Gulf. The whole flight had taken less than seven minutes. There were no survivors.

[¶48.] By noon that day, Iranian helicopters and boats began to search the area and recover the bodies. It was not until later in the day that the officers and men of the Vincennes would learn that what they had shot down was not an Iranian F-14, but a commercial, civil flight.

[¶49.]

Command, Confusion, and Cognitive Locking

[¶50.] Since the "black box" flight recorder on board the Iranian Airbus has been irrecoverably lost in the waters of the Persian Gulf, we shall never know exactly what her flight profile was, whether the crew ignored the American challenges or simply did not hear them, or whether they were aware of the fight going on beneath them or the impending attack.

[¶51.] However, the Vincennes had a black box of its own. The SPY-1A, Command and Decision, and Weapons Control System computers were all equipped with magnetic tape equipment that tracked and recorded all of the signals received and processed by these key pieces of electronic equipment. ¹⁷ Because of this, we have been able not only to verify the timing and nature of all of her actions, but also to obtain a remarkable, perhaps unique, cross-check of the way that scenario preconceptions can distort both perceptions and memory.

[¶52.] The situation aboard the Vincennes that day was one of confusion and disorder. This harsh judgment is borne out by the transcript of the Navy's official investigation report and, in retrospect, by the divergence between the memories and oral testimony of the officers present and the actual, recorded sequence of events. ¹⁸

[¶53.] The story told by the data tapes is straightforward. Iran Air Flight 655 took off from Bandar Abbas at 10:17 a.m. on the morning of July 3, on a heading of 210 (runway 21). Squawking Mode III, Code 6760 continuously, it kept on a more or less constant heading of 210, climbing steadily to its cruising altitude while gradually gaining speed. Data and testimony from the USS Sides corroborate the flight path and the Mode III IFF squawk. Indeed, the Sides was to identify the unknown aircraft as nonhostile and turn its attention elsewhere only seconds before the Vincennes launched its missiles. ¹⁹

[¶54.] The story told by those inside the CIC aboard the Vincennes is quite different. From the first alerted contact, various personnel began to report a "Mode II" squawk on a code associated with Iranian F-14s. Although none of the data recorders reported any IFF response other than Mode III, Code 6760, those aboard the Vincennes continued to consistently misreport the signal.

[¶55.] As the range closed, the Vincennes began to broadcast increasingly urgent warning messages to the unknown aircraft; at first, these were general challenges on both military and international civil distress nets. But as the notion that the aircraft was indeed an F-14 became fixed in the minds of the key operators, the challenges were made more specific and were addressed only to an unidentified "Iranian F-14." A quick thumb-through of a listing of commercial flights missed the clear listing for Flight 655, although it was on course and nearly on time.

[¶56.] A warning of possible "COMAIR" (commercial aircraft) issued a minute or two later was acknowledged by the CO, but essentially ignored. At this point, the ship was still engaging the Iranian surface boats. Moreover, the ship was heeling sharply with loose books and equipment flying about the CIC. With the TAO concentrating on the surface battle and his attention divided, and the Anti-Air Warfare Commander (AAWC) new to his post (and generally regarded as inexperienced and a weak leader), de facto leadership fell upon the more junior Tactical Information Coordinator (TIC), who by that time was almost literally shouting about the immediacy and seriousness of the threat.

[¶57.] To give Captain Rogers credit, he did allow the unknown aircraft to close to well within its possible missile firing range before asking for and receiving permission to intercept, and he did so only after repeating the challenge several more times. Only then, convinced that the threat to his ship was too serious to ignore, and under pressure to act quickly to avoid the earlier fate of the USS Stark , did he authorize the firing.

[¶58.] Was he justified in his perception of a real threat to his ship (which was the Navy's claim)? Were the Iranians reckless for flying over a fire-fight in progress (which they may have been-- if those at Bandar Abbas were aware of the exact position of the fight)? Was the whole incident a regrettable, but unavoidable, accident of war (which is precisely what the resulting U.S. attitude was, in the Pentagon, in Congress, and in the press)? Possibly all three of these U.S. assertions are true. But are they relevant?

[¶59.] The first question to be asked is: Was an error made on the U.S. side at all? The U.S. Navy finally claimed that Captain Rogers of the Vincennes acted correctly in appraising the threat, and ultimately awarded him a medal for his performance in the Gulf that day. ²⁰ Others in the United States asserted that such blame as there was attached solely to Iran. ²¹ Iran, on the other hand, went so far as to claim that the United States had shot the Airbus down in full knowledge that it was a civil aircraft, which of course would have been an evil deed rather than an error.

[¶60.] It would be better to follow Lanir's example and ask whether there was an error made at all. Complex military-technological systems such as those aboard the Vincennes , designed to reduce the incidence of human errors, also provide mechanisms for producing new ones. The question, then, is not whether any individual could have acted correctly in theory, but how and why the socio-technical command-and-control system contributed to the consequential failure that occurred.

[¶61.]

The Investigations

[¶62.] The large-scale technical military system operating in the Persian Gulf that day, of which the Vincennes was the central feature, was not waging total war, but rather a highly selective engagement in an arena known to be filled with civil traffic on air and sea. This very sophisticated piece of equipment had been placed in a situation for which it had never been designed precisely because it was thought to be most capable of making the kinds of quick and accurate judgments that would be necessary. It failed.

[¶63.] Of course, the Vincennes itself is a piece of machinery, and only human beings can make "errors." Once they had ascertained that the equipment had operated properly, the only question for the investigators was whether the source was at the personal level (i.e., the performance of the CO or the TAO) or the group level (i.e., collective failure in the CIC). It was this general approach that framed the subsequent investigations. ²²

[¶64.] Navy hearing boards such as the one convened to review the events of July 3, 1988, are unique in focusing on the CO, owing to historical naval tradition that it is the CO, and only the CO, who is responsible for anything and everything that happens on or to his ship. As often as not, judgment calls are made on the simple basis of what the hear-ing officers would have done in the same situation, given the same information.

[¶65.] This tradition worked against the possibility of a comprehensive, systemic investigation into the circumstances preceding the missile firing. ²³ For the question should not have been whether the CO was justified in taking the actions he did given the situation and the information he had, but how the situation had developed so badly and why the information being provided was so skewed from reality. These matters were in fact addressed by the investigation, but by no means to the degree or depth that would have been required to develop an adequate set of answers.

[¶66.] The investigation board was convened by Rear Admiral William M. Fogarty at Bahrain beginning on July 6, while the events were still fresh in the minds of the participants. Formal hearings began a week later, and the entire procedure was completed and the report delivered to the Navy on July 28. ²⁴ Even in the cleansed form provided to the public, the report is rich in personal and technical detail. Perhaps the most striking feature is the degree to which the recollections of the participants as to the nature and assessment of the presumptive threat differ, and the variance between what was reported by the SPY-1A computers and what its human interpreters were reporting.

[¶67.] The record shows that the decision to fire was taken more or less calmly and deliberately on the basis of personal advice passed from junior officers to the senior AAWC, and from the AAWC to the CO--in the face of a stream of contrary evidence from the electronics aboard. Confronted with the problem of reconciling the manifest mistakes made in interpretation of technical evidence, the board concluded that "stress, task-fixation, and unconscious distortion of data may have played a major role in this incident." ²⁵ The report then went on to attribute the distortion to the relatively junior TIC and Identification Supervisor (IDS), who became convinced that the track of Iran Air 655 was an F-14 after an IDS report of a momentary Mode II squawk.

[¶68.] The Fogarty report states: "After this report of the Mode II, TIC appears to have distorted data flow in an unconscious attempt to make available evidence fit a preconceived scenario (`scenario fulfillment')." This fulfillment including continuing to read the Iran Air flight as descending toward the Vincennes , even though the information being presented by the electronic suite was contradictory. In such circumstances, it may indeed be remarkable that the CO, deprived of any direct source of information or data, was able to delay his decision so long.

[¶69.] Whether or not he would have been justified to shoot if the evidence were simply ambiguous or the uncertainties known is not the center of our concern here. Rather, the focus of my analysis is the manifest failure of the decision-making system on the Vincennes to interpret the evidence correctly. The system "failed" in that a false hypothesis was constructed and maintained, serving as the basis for all subsequent actions. This would have been as serious a matter even if Captain Rogers had decided not to fire, although it is most likely that we would never have learned about it. Among the conclusions of the Fogarty report is a general recommendation:

[¶70.]

Since it appears that combat-induced stress on personnel may have played a significant role in this incident, it is recommended the CNO (Chief of Naval Operations) direct further study be undertaken into the stress factors impacting on personnel in modern warships with highly sophisticated command, control, communications, and intelligence systems, such as AEGIS. This study should also address the possibility of establishing a psychological profile for personnel who must function in this environment.

[¶71.] Stripped of its deliberately restrained prose, this is a quite remarkable admission that the very sophistication of the system may in itself have been a contributory factor. This is the only statement in the official report to acknowledge that it might be wise to differentiate between questions of personnel performance and training at the individual and/or group/ship level and more systemic factors.

[¶72.] Nevertheless, and following U.S. Navy tradition, the entire review process was treated as a search for "culpability." Having absolved the individuals of blame, the Navy, as well as the press and Congress, then moved rapidly to the highest possible level of political analysis; if there was no personal malfeasance, and the Aegis system had worked perfectly, the result could only be attributed to the misfortunes of war.

[¶73.] On October 6, 1988, a panel of five psychologists chosen through the American Psychological Association testified at a hearing of the House Armed Services Committee. ²⁶ As the Fogarty report had done, the expert panel pointed out that in an era of increasing technical complexification, it will no longer do to continue to point only to "operator error" as the source of malfunctions and disasters. ²⁷ Rather, what happened aboard the Vincennes on July 3 could be seen as part of what one psychologist characterized as the "glass cockpit" syndrome discussed in chapter 7, the result of humans operating large-scale, highly sophisticated technical systems under conditions of high stress, high consequence, and high visibility. ²⁸

[¶74.] Regrettably, the congressional hearing panel was more interested in the question of the functioning of the Vincennes electronic hardware (and software) than in its human interface. Much attention was paid to whether the very expensive and highly sophisticated Aegis radar system did work, as a technical system. In fact, it did work--in the sense of identifying and correctly tracking Flight 655 almost from the moment of its takeoff--although it did not, and could not, given its technical limitations, identify the type of aircraft, or even its size. The rest of the electronics suite also worked well. As with Lanir's case, the "surprise" was the degree to which the system lent itself to misinterpretation and mindset by its operators.

[¶75.] The Vincennes failed as an operating organization, however perfectly its equipment was operating. ²⁹ Yet, the emphasis in the official investigation, and in the reports of the five APA psychologists, was on the role played by stress. "Scenario fixation," in which the Iran Air flight became embedded in the collective consciousness of those in the CIC as a hostile Iranian F-14, was derived from and attributed to the degree of stress present rather than to the human-machine interface.

[¶76.] The testimony was almost unanimous in stating that stress in CIC was very high, although there is no way retroactively to determine just how high it was, or whether the sense of near-panic so clearly evinced by the TIC was spreading. But what are the expectations of a combat system such as the Vincennes , or of the CIC as a war-fighting center? That stress would be low? That battle conditions would be other than confusing? That the ship could be attacked on the surface, or from the air, but not both simultaneously--not to mention possible subsurface attacks in other circumstances? If these are or were the assumptions under which the Aegis cruisers were designed, than the Vincennes should never have been deployed into the Gulf.

[¶77.] Stress was clearly a contributing factor, but its presence is in no way explanatory, and by no means exculpatory. Stress is, or should be, a central assumption of battle, along with confusion, uncertainty, and risk. To design military systems otherwise would be folly. ³⁰ Looking back over naval history, there is no evidence that the Vincennes was under a degree of stress greater than one would expect for a ship in combat, in strange waters. If the ship as a system is incapable of operating correctly under such circumstances, she is a failure as a weapons system, however well her machinery, electronics, and missiles perform. ³¹

[¶78.] The Navy sought to excuse the Vincennes 's performance on the grounds that the technology worked as designed, but the imperfect human beings did not. But these are empty excuses. Any real-world technology must be designed around the fundamental premise that humans are imperfect. Neither the unanticipated combat situation, nor the degree of stress that was manifest, nor the resulting confusion that occurred aboard the Vincennes can excuse her poor performance. Military systems that cannot function under stress, or amid confusion, or while under multithreat attack, are useless in the real world. ³²

[¶79.]

Conclusion

[¶80.] Taken as isolated events, these three cases may seem to be no more than another tragedy of armed conflict. In a tense situation in a combat zone, an unfortunate mistake caused the deaths of several scores or hundreds of people. It has happened in every war, and will happen again in the next. There appears to have been little interest in it as a case of malfunctioning technology. As a purely political story, it was of continuing interest for a while because of the subsequent attempts by both countries to assign culpability, and perhaps deliberate malfeasance to the other. ³³

[¶81.] When similar errors occur in the fog and confusion of real war, it is often difficult to disentangle the circumstances. But these three incidents occurred in relative isolation, and under relatively clear and uncomplicated circumstances. What they have in common is that each represents a case where the system failed even though the equipment worked perfectly. In each case, the ultimate cause was more than just cognitive failure or undue stress. It was a failure of representation based primarily on the difficulty of forming an independent inter-pretation of the situation in the environment created by the new equipment.

[¶82.] The Israeli aircraft had no interest in shooting down an unarmed airliner, but the command-and-control system had shaped a decision that it was a hostile aircraft. The attack on the Stark was probably not deliberate, but rather the consequence of quick-response high-tech warfare, and the Stark itself was vulnerable because its automated quick-response defenses were considered potentially threatening to other aircraft in a presumably moderate threat environment. The very sophisticated, high-technology Aegis aircraft detection and track identification system that is the heart of the Vincennes , designed to discriminate between U.S. and hostile aircraft at high speed in the heat of combat, provided a frame in which quick response was the dominant concern, providing neither space nor time for reflection on the nature of the putative threat.

[¶83.] Because the United States has been heavily investing in the kind of sophisticated, computer-aided and controlled technological military systems these examples represent, and will be relying even more heavily on them in the future, it is doubly unfortunate that the official investigations left so many avenues unexplored. But that has been the usual situation in almost every similar circumstance, civilian or military, despite a growing body of literature on systemic failure in complex, high-technology organizations. ³⁴ The most durable pocket of military tradition is the belief that human beings are ultimately responsible and ultimately in control, until and unless their equipment fails them.

[¶84.] In her extensive study of the effects of automation and computerized equipment in a variety of civilian work settings, Zuboff has noted that the transition to a more computerized workplace is also accompanied by a transfer of emphasis from oral-culture oriented, action-centered skills to reflective, intellective ones. In contrasting an older and still traditional plant (Piney Wood) to a newer and highly automated one, one worker summed up the difference in reaction to problems and difficulties this way: "When there is a problem at Piney Wood, someone goes out and kicks something; when there is a problem at Cedar Bluff, the operators have a meeting." ³⁵ Clearly, when the human in the loop is not fully a part of it, what is most desperately wanted is some reflective and digestive time. This raises serious questions about the wisdom, or effectiveness, of inserting highly automated equipment into an environment that is by tradition very much human-oriented and action-centered. ³⁶

[¶85.] That the overall effect of the new technologies might be a situation where people acting as expected with equipment that is functioning perfectly will fail to arrive at the objectively correct interpretation is rarely, if ever, considered. The Navy hearing boards never considered the possibility that the failure lay in the theory of design, that a highly automated, rapid-response battle system that depends for its function on real-time interactions between complex computerized systems and human operators may have an inherently high probability of error in any crisis situation that has not been anticipated, planned, and rehearsed.

[¶86.] Because fundamental surprise has long been identified as the essence of warfare, the prognosis for the future is frightening. But as will be discussed in the following chapters, this has not stopped the military from moving ever more rapidly to extend computer control and automation onto every aspect of the battlefield, from missile defense in outer space to fire control for the grunt in the foxhole.

NOTES:

1 The complete story of the Wampanoag is wonderfully told by Morison, Men, Machines and Modern Times .

2 Ibid., 111ff.

3 Douglas, "Navy Adopts the Radio," 131.

4 Deitchman, Advance of Technology , 244-245.

5 There is a long list of books on technical innovation in the military, of which those of Deitchman, Advance of Technology , Rosen, Winning the Next War , Smith, Military Enterprise and Technological Change , and van Creveld, Technology and War , are of particular interest because of their multiple case studies and comparative sensitivity to organizational issues.

6 Morison, Men, Machines and Modern Times , 144. The U.S. and other navies still maintain a few of the tall ships for training cadets. Because of the special problems of command at sea, navies are in many ways the most traditional of military services, and the most self-conscious about their culture.

7 Morison, Men, Machines and Modern Times , 117.

8 Deitchman, Advance of Technology , 245.

9 The phrase comes from Zvi Lanir, who, along with Chris C. Demchak, contributed greatly to my thinking about computers and cognition in the military. The details that follow are taken from Lanir, "Reasonable Choice of Disaster."

10 Lanir points out that the investigators were fortunate in having access to all radio transmissions, Egyptian and Israeli, as well as the cockpit flight recorders. Therefore, it was not so easy to follow the usual practice of seeking to assign causality to a single human action or a single point of failure, as was the case with the shooting down of Korean Air Lines Flight 007 over Kamchatka. Although more details about the failures of Soviet air defenses are now being revealed, arguments about cause raged for many years.

11 This is very similar to Barry Turner's observation that human societies are inherently "negentropic," and therefore capable of self-organizing disasters out of ordinary conditions and circumstances. See, for example, Turner, Man-Made Disasters .

12 Because the United States was not at that time formally at war with Iraq, and regarded Iran as the more hostile to its interests, AWACS were in the habit of reporting Iraqi aircraft as potential strike support rather than hostile--a decision reinforced by occasional overflights of U.S. ships by Iraqi aircraft in the past.

13 U.S. Congress. House of Representatives, Committee on Armed Services, Iraqi Attack on the U.S.S. Stark .

14 U.S. Congress, Iraqi Attack on the U.S.S. Stark . For an excellent, concise description, see Vlahos, " Stark Report."

15 This section is based on Rochlin, "Iran Air Flight 655."

16 Most of the detail in this section is obtained directly from Fogarty, Formal Investigation into the Circumstances Surrounding the Downing of a Commercial Airliner by the U.S.S. Vincennes (CG49) on July 3, 1988 , hereinafter referred to as the "Fogarty report."

17 The Aegis Display System data could not be extracted, precluding any positive confirmation of actions taken at the CO and TAO consoles.

18 This was much clearer in the perusal of the Fogarty report than it was in subsequent news stories.

19 To quote the Fogarty report directly: "The data from USS Vincennes tapes, information from USS Sides and reliable intelligence information, corroborate the fact that [Iran Air Flight 655] was on a normal commercial air flight plan profile, in the assigned airway, squawking Mode III 6760, on a continuous ascent in altitude from takeoff at Bandar Abbas to shoot-down."

20 Washington Post , 30 April 1990. The Legion of Merit, the U.S. armed forces second highest award, was presented to Captain Rogers and Lieutenant Commander Lustig (the weapons officer) for meritorious conduct and "heroic achievement" on July 3, 1988. The citations did not mention the downing of the Iran Air flight at all.

21 Admiral Crowe was reported to have concluded that airport officials at Bandar Abbas "should not have allowed the flight to take off while a firefight was going on some miles away" ( New York Times , August 20, 1988). Then Vice President George Bush was quoted as saying: "I will never apologize for the United States--I don't care what the facts are." For these and similar quotes, see, e.g., George Wilson, "The Risks of Shooting First" ( Washington Post , January 28, 1989). It should, however, be noted that the Navy's conclusion that Captain Rogers was justified in firing in presumptive self-defense was widely shared on Capitol Hill. See, e.g., Congressional Quarterly , July 9, 1988, 1905ff.

22 It is rare that such investigations touch on the broader and more systemic level in which the actors, individually and collectively, were placed in a situation in which the probability that they could exercise proper judgment was greatly reduced. See, for example, Weick, "Mann Gulch Disaster."

23 This arises from the usually praiseworthy desire to expedite inquiry while evidence is fresh and untainted. The problem lies not with the speed of the first inquiry, but with the closing of the book on the entire incident once this hurried and summary court, held in the combat area, far from other sources of information and analysis, is concluded.

24 The version of the Fogarty report released to the press on August 19 was cleansed of classified material, and of potentially damaging references to specific individuals.

25 Fogarty report, 45.

26 Baruch Fischoff, Robert Helmreich, Richard Nisbett, Richard Pew, and Paul Slovic each presented testimony "on behalf of the American Psychological Association." These are summarized in U.S. Congress. Senate, Committee on Armed Services, Downing of an Iranian Airliner by the U.S.S. Vincennes .

27 Testimony of Baruch Fischoff before the Committee on Armed Services, U.S. House of Representatives, on behalf of the American Psychological Association, October 6, 1988.

28 Testimony of Robert L. Helmreich before the Committee on Armed Services, U.S. House of Representatives, on behalf of the American Psychological Association, October 6, 1988. Also see Squires, "Glass Cockpit Syndrome."

29 See, for example, Perrow, Complex Organizations . This crucial distinction is rarely made in the policy debate over the military and its equipment, or even for analogous systems such as nuclear power plants or chemical refineries.

30 Unfortunately, Aegis cruisers may well have been so designed. The CIC is located high up in the superstructure, behind the Aegis radar panels rather than in the small and already cramped main hull. Aegis cruisers, with their light displacement and narrow beam, roll easily, and the CIC location multiplies the effects of the ship's motion on those fighting it. Their primary mission is that of long-range air defense of an aircraft carrier Battle Group in blue-water operations. Thus, they represent yet another move in naval warfare away from the close grappling of ancient days through cannon, rifled guns, and aircraft to remote combat fought at extreme ranges by electronics.

31 Byron, "Surface Navy."

32 In principle, this is also the position of the Department of Defense. Yet the U.S. General Accounting Office found that the DoD was generally not complying with its own Directive 5000.3 requirement that testing should be conducted under conditions realistically simulating combat stress. See, for example, U.S. General Accounting Office, Weapons Testing .

33 See, for example, Jeff Cohen and Normal Solomon, "Today's News Next Year, Maybe," San Francisco Examiner , July 12, 1992, A11.

34 Demchak, Military Organizations ; Roberts, New Challenges ; Rochlin and others, "Self-Designing High-Reliability Organization"; La Porte and Consolini, "Working in Practice but Not in Theory"; Perrow, Normal Accidents ; Sagan, Limits of Safety .

35 Zuboff, Age of the Smart Machine , 196.

36 The match between oral culture and action orientation has also been documented in naval flight operations. See Rochlin and others, "Self-Designing High-Reliability Organization"; Rochlin, "Informal Organizational Networking."

Jul	SEP	Aug
	05
2004	2006	2009