The United States entered World War I almost completely unprepared: the National Defense Act which Congress had passed in 1916 had provided the basis of a mobilization plan, not an actual army. In early 1917 the country had only 210,000 men under arms, a third of them National Guardsmen who had been called up the previous summer to serve on the Mexican border. The Army had no permanent tactical organization above the level of the regiment and lacked adequate quantities of artillery, machine guns, tanks, modern aircraft, and even gas masks. Its General Staff organization was not designed to cope with the logistical and operational problems presented by a major conflict, and at the direction of the Wilson administration it had made no war plans. The Army had no intelligence organization.
Within seventeen months, however, the country had transformed itself into a fighting machine. With the help of the draft, the United States raised an Army of 4 million men; half of this great force was transported to France, where it provided the decisive margin that led to victory over Imperial Germany and its allies. American industry was also mobilized for war, but not soon enough, forcing Britain and France to supply the American Expeditionary Forces (AEF) in France with almost all required tanks, planes, and artillery. The General Staff was repeatedly restructured and finally became an effective instrument of control. And under the pressures of war, the Army was forced to develop new intelligence capabilities.
The development of Army Intelligence in World War I proceeded along two parallel but more or less separate tracks. At the War Department level, intelligence work was revived; by the end of the war, Military Intelligence had become a fullfledged staff element as one of four operating divisions of the General Staff. Concurrently General John J. Pershing, commander of the AEF built up his own field intelligence organization, structured along rather different lines to meet tactical needs. Pershing's ideas on Army reorganization and on the need for an intelligence apparatus were adopted by the War Department in organizing troop units and ultimately provided the basis for reorganization of the General Staff itself.
22
Van Deman and Military Intelligence
At the War Department level, the individual most responsible for rebuilding an intelligence apparatus was Maj. Ralph Van Deman, the General Staff officer who had lobbied vigorously but unproductively in the years before the war to re-create the Army's capabilities in this area. In 1917, with the United States at war, Van Deman tried again. He submitted a formal recommendation to Maj. Gen. Hugh Scott, the Army's chief of staff, urging the War Department to establish a Military Intelligence Division. This initiative was strongly supported by Van Deman's immediate superior, Brig. Gen. Joseph Kuhn, chief of the War College Division. 1 Unfortunately, the chief of staff was not receptive. Scott was an old-time cavalryman with an encyclopedic knowledge of Indian sign languages and a deep interest in packsaddles, but his former commander in chief, President William Howard Taft, had described his intellectual abilities as "wood to the middle of his head." By 1917 he was demonstrating a disconcerting tendency to fall asleep at Cabinet meetings.2 Scott simply believed that the Army had no need for an intelligence organization: since America was now fighting a war as an ally of Great Britain and France, the United States could acquire whatever intelligence it needed from them.
Undeterred, Van Deman decided to adopt different tactics, employing what British military analyst Basil Liddell Hart would later describe in another context as the "indirect approach."3 Using such diverse contacts as a lady novelist and the police chief of Washington, D.C., as intermediaries, Van Deman discreetly lobbied Secretary of War Newton D. Baker. Baker proved to be more receptive to the major's ideas, and on 3 May 1917, the War College Division replaced its moribund Military Information Committee with the Military Information Section headed by Van Deman.4 Although far from an ideal organizational setup, with Military Intelligence subordinated to what was essentially a planning organization, it was at least a beginning.
The Military Intelligence Section began modestly, with Van Deman, two other officers (one retired), and two civilian clerks. Initially, its office space was confined to a balcony overlooking the War College Division's library. At first it had no data files, since the information collected by the old Military Information Division had been merged into the War College Division's general files and remained there, effectively lost. Its responsibilities, however, were considerable. In addition to supervising the existing Army attache system and
23
developing policies and plans for Army Intelligence activities, Van Deman's organization was charged with "the supervision and control of such system of military espionage and counterespionage as shall be established . . . during the continuance of the present war."5 The section was thus not only to serve as a staff element but was also to perform operating functions.
In the beginning, Van Deman's organization depended heavily on the British and the French, who supplied it with watch lists of suspected enemy sympathizers and basic organizational concepts. From the British, Van Deman borrowed the fundamental division of intelligence into "positive" and "negative." Positive intelligence consisted of collecting information from the enemy; negative intelligence consisted of denying the enemy intelligence about one's own forces. An important part of negative intelligence was "counterespionage," another new word in the Army's growing intelligence vocabulary, this one borrowed from the French.
Doctrine and data could be borrowed from abroad, but personnel could not. Because of the demand for regular officers in combat assignments, no more than six regulars were available for intelligence duties at the General Staff during the whole war. To obtain additional competent personnel for his own organization, Van Deman used an informal "old boys" network that recommended civilians for direct commissions, a practice not uncommon in the World War I Army Initially, many of these men were commissioned in the Signal Corps, since there were unused officer billets in the Aviation Section of that branch. Later, intelligence officers were hastily commissioned in whatever Army branch could afford to give up a few of its slots.
The buildup of the Military Intelligence element of the War Department General Staff took place at the same time that hundreds of additional intelligence officers had to be found for the Army's divisions and subordinate units. In July 1917 Pershing and the War Department agreed that intelligence staffs had to be provided to units down to the battalion level. Logically, this was an area in which the General Staff's intelligence organization should have had an interest. However, the Military Intelligence Section was too small to supervise the procurement and training of intelligence personnel for the field army instead, this was accomplished through normal Army channels. The process began in July 1917, when the adjutant general directed departmental commanders to select 160 specially qualified men from the newly established officer training camps for intelligence assignments. The ideal individual, the adjutant general advised, was a "young college instructor" with language ability. 6 The Signal Corps trained intercept operators and photo interpreters, and the Corps of Engineers trained topographic personnel. The Military Intelligence Section did process
24
applications for the Corps of interpreters, a new Army organization of officers and noncommissioned officers set up in July 1917 to handle the Army's language functions. Later, it procured enlisted counterintelligence specialists.
Intelligence Operations
The General Staff's former Second Division had confined its efforts to gathering positive intelligence, employing the Army's military attache system to collect scientific, technical, and geographic data from abroad. Although the Military Intelligence Section was responsible for the lapsed functions of the older organization, it got off to a curiously slow start in collecting foreign intelligence. Until the end of 1917, only one officer was assigned to the task, and the principal target was Mexico. However, as his organization enlarged Van Deman was able to direct greater resources to this area. In doing so, Military Intelligence gained a broader definition. In addition to accumulating data on the military situation abroad, the Army began to collect information on economic, social, political, and even psychological factors worldwide. World War I was a global conflict fought between entire industrialized societies, where victory depended on more than military factors; the fighting armies were only the cutting edges of much larger swords. The War Department's Military Intelligence staff element thus became the functional equivalent of today's Central Intelligence Agency.
However, the War Department's efforts at collecting foreign intelligence were somewhat overshadowed by the fact that the AEF's general headquarters (GHQ) in France was 3,000 miles nearer the enemy and in a much better position to gather information on the European theater. Pershing had already established his own intelligence staff element, and this functioned almost autonomously The great distances between the two separate organizations discouraged collaboration, as did Pershing's belief that he was responsible for the conduct of the war and that the only function of the War Department was to furnish him with the troops and supplies he needed.
Because of this, Van Deman directed much of his attention to the new field of negative intelligence, or counterintelligence. 7 This was an area in which the Army had little previous experience. Although the Army had conducted counterintelligence work during the Philippine insurrection, not since the Civil War had it contended with the problems of espionage, sabotage, or subversion in the continental United States. However, when the United States joined the war against Germany in 1917, it appeared that the country confronted a substantial threat from within. The America that entered World War I was still a nation of
25
immigrants, many newly arrived. German Americans were particularly suspect, but the War Department was also concerned about the loyalties of Irish Americans, Scandinavian Americans, and African Americans. In addition to the problems that might be posed by unassimilated ethnics, there was a substantial antiwar movement. Finally, opinion-makers at the outbreak of war had exaggerated ideas about the scope and power of the German espionage and sabotage organization within the country.
In 1917 the United States seemed almost defenseless against these perceived enemies from within. The Treasury Department had a Secret Service, but it was confined by law to narrowly circumscribed duties. The Department of justice maintained the Bureau of Investigation, but the bureau's duties before the outbreak of war had largely consisted of investigating cases of fraud against the government. A few major cities had organized police "bomb squads" to deal with the anarchist threat of the period. The almost total lack of civilian resources in the field spurred the Army to launch its own major and widesweeping counterintelligence program.
In June 1917 Van Deman took his first step, setting up a War Department security force of civilian investigators drawn from the ranks of the New York Police Department's Neutrality and Bomb Squad. Its operations were cloaked in secrecy, with members working from a private office building in Washington, D.C., under the enigmatic designation "Personnel Improvement Bureau." At first intended as a guard force, the unit soon began screening military personnel and applicants for government employment. A month later, the Military Intelligence Section opened its first field office in New York City, also staffed by former New York City policemen. Six additional field offices were subsequently set up in other major cities and embarkation points to provide counterintelligence coverage.
This was only the beginning. The Military Intelligence Section was concerned particularly with the problem of possible subversion within the vast new citizen forces being raised by the draft. The draft act which Congress passed in May 1917 had been designed to tap as much of the national manpower pool as possible. Granting few exemptions, the act impartially swept up American citizens and resident foreign nationals, including citizens of enemy countries. Regarding this heterogeneous force as posing a serious threat to national security, Van Deman believed that the newly forming National Guard and National Army divisions were infested with German agents and sympathizers.8 In October 1917 he ordered the divisional intelligence officers just assigned to these units to come to Washington, D.C., under tight security. Upon reporting, the officers were instructed to set up a secret surveillance program within their
26
divisions. The program was later extended to Regular Army divisions and fixed installations with the assistance of a confidential pamphlet, "Provisional Counter-Espionage Instructions," drawn up by Van Deman's staff.
Van Deman conceived a comprehensive counterespionage program. It envisaged the creation of a clandestine agent network extending throughout the Army down to company level. Nets in each division would be managed by an assistant to the divisional intelligence officer. He would work through a system of anonymous collection managers known only to himself and to their own immediate superiors and subordinates within the apparatus. At the bottom of this secret pyramid, "operatives" placed in every company would submit intelligence reports on their fellow soldiers. At least two operatives, mutually unknown to one another, would be recruited from each company.9 Reports emanating from this organization would be relayed by the divisional intelligence officer to the Military Intelligence staff in Washington for investigation.
Once this system was in place, it produced a growing stream of incident reports that drove the relentless expansion of the War Department's counterintelligence organization. To supplement the efforts of his overextended force of intelligence officers, Van Deman hired additional civilian detectives, recruited unpaid volunteers, and soon found a fresh source of investigative manpower in the enlisted counterintelligence specialists of a completely new military organization, the Corps of Intelligence Police.
The Corps of Intelligence Police (CIP)
The stimulus for the creation of this corps of enlisted investigators had originally come from the AEF Soon after his arrival in Europe, Maj. Dennis E. Nolan, Pershing's intelligence officer, had become concerned about the possible security problems faced by American troops fighting on foreign soil. In 1915 Nolan had written alarmingly about the possibilities of German invasion; now his work took on a more concrete form. In early July 1917 he requested that the adjutant general provide him fifty company grade officers proficient in foreign languages. In addition, Nolan asked for "fifty secret service men who have had training in police work [and] speak French fluently."10 This called for a category of intelligence personnel not previously imagined by anybody in the War Department, but in August the acting chief of staff authorized the creation of a fifty-man Corps of Intelligence Police made up of enlisted soldiers who would serve with the "rank, pay, and allowances" of sergeants of infantry.11
27
Tasked with furnishing the appropriate personnel, Van Deman ran into difficulties. Private detective agencies seemed a likely source at first, but when told of the Army's requirements for French-speaking investigators, the head of the Pinkerton Agency countered, "There ain't no such animal."12 The War Department was reduced to recruiting the first CIP agents through newspaper advertisements. The first contingent was assembled at Fort Jay in New York Harbor, given a month's training as infantry, outfitted with distinctive greencorded campaign hats, and shipped to France without civilian clothes or any instruction in intelligence work. Once overseas, the group was screened by French authorities, who rejected many as undesirables. Those who passed muster were then bustled off to Le Havre for instruction by veteran Allied counterintelligence officers. 13
The initial group was not promising, but the need for enlisted counterintelligence specialists remained, and the formation of the CIP had set a precedent. The Military Intelligence Section found it increasingly difficult to staff its headquarters and expanding network of field offices with the existing mix of freshly commissioned reserve officers, civilian volunteers, and hired detectives. Officers were a scarce commodity, and competent civilian investigators were hard to find, especially since the War Department paid them only $4 a day plus expenses. Thus it seemed logical for Van Deman's organization to turn to the enlisted ranks of the Army to solve the personnel problems in the counterintelligence arena. The iron broom of the draft had swept highly qualified people into the ranks of the Army; men experienced in law, teaching, or insurance investigation were especially fit for counterintelligence work. Additionally, the military thought that investigations of Army personnel could be carried out most appropriately by other soldiers. In November the Military Intelligence Section requested that it be allotted 250 CIP agents to assist its counterintelligence program. Many of the civilians previously employed by Van Deman promptly enlisted in the Corps of Intelligence Police, including the twenty-three former policemen in the New York Field Office.
Expansion
The expansion of Van Deman's organization, largely driven by operational responsibilities in the counterintelligence field, in turn forced a growing specialization within the Military Intelligence staff. It was no longer possible for Van Deman and a small group of assistants to deal interchangeably with all aspects of positive and negative intelligence. The whole operation had to be put under a
28
bureaucratic regimen. Between December 1917 and January 1918, Van Deman divided his organization into functional subsections, all according to the British practice with which he was now familiar. As the organization grew larger and more complex, it achieved a position of greater prominence within the War Department. In February, following a reorganization of the General Staff, the Military Intelligence Section was upgraded in status, becoming a branch of the newly established Executive Division. Since the Executive Division was burdened with diverse responsibilities, the reorganization was less than perfect. As one knowledgeable officer put it, its chief "could not know and would not know what MI was doing . . . when MI papers came up to him, they were like Greek to him."14 In March the Military Intelligence Branch, overcrowded at the War College, moved to a seven-story apartment building in downtown Washington, D.C., where it was at least in closer proximity to the Army's center of administrative power.
The Military Intelligence Branch initially consisted of eight numbered sections. MI-1, MI-5, MI-6, and MI-7 were responsible for carrying out general support functions in the respective areas of administration, publications, translation, and management of confidential files. Later, MI-7 assumed responsibility for graphics. At first there was only one section, MI-2, exclusively dedicated to collating foreign intelligence, although MI-5 soon received the new mission of coordinating the collection efforts of the attaches. Two other sections were set up to manage aspects of counterintelligence: MI-3, which handled counterespionage in the military services, and MI-4, which dealt with civilian subversion.15 MI-3 worked closely with over 400 divisional and installation intelligence officers who supervised the clandestine counterespionage system. Other subelements served specialized needs: administering the District of Columbia field office, dealing with the specific counterintelligence problems presented by foreign-bom draftees, and overseeing programs that dealt with particularly sensitive Army branches such as the Air Service and the Chemical Corps. Thousands of investigations were conducted, but contrary to early fears, only a relatively small number of troops had to be removed. The draft-raised Army was loyal.
In addition to policing the Army's own ranks, Van Deman was concerned that spies and agitators in the civilian community might also threaten the Army or at least its mobilization base. At the same time that MI-3 was created, MI-4 came into existence to handle counterintelligence in the civilian sector. This section sought to cope with broad and illdefined threats, initially concerning itself with labor unrest in the West, racial disturbances in the South and Southwest, and foreign disaffection in the polyglot cities of the East. Its operative premise
29
was that "the misbehavior, disloyalty, or indifference of native Americans is as important a material of military intelligence as any other."16 At one time or another, the organization involved itself in deportation cases, sabotage by organized labor, enemy finance and trade, and counterespionage work abroad.
To carry out its duties, MI-4 had to conduct an active liaison with many other government agencies, especially the Department of justice, which alone had powers of arrest and prosecution in cases of civilian offenses against the military. MI-4 also relied on two civilian auxiliaries, the Plant Protective Service and the American Protective League. The Plant Protective Service was an organization of undercover civilian operatives originally established by the chief signal officer to protect the country's new aircraft industry. Subsequently it spread to other private plants working under government contract. Almost inevitably, the organization came under the aegis of Military Intelligence.
The American Protective League, larger and less official in character, comprised several vigilante groups originally put together to help the Department of justice uncover spies. Once in place, it also began to assist local authorities in enforcing the draft act, and its activities brought it into a close working relationship with MI-4. The Army sought to curb the organization's excesses while still making use of its thousands of members: estimates of the league's strength ranged between 60,000 and 200,000, with the organization itself favoring the latter number. Ultimately, top American Protective League leaders were commissioned as officers in MI-4 to enhance military control over the league's activities. Although such paramilitary organizations appear alien to the American tradition, similar organizations existed in Great Britain and France, reflecting the intense and sometimes excessive nationalism of the period.
In addition to his pioneering work in establishing an Army counterintelligence organization, Van Deman also involved Army Intelligence in the exotic world of codes and ciphers. This was an area in which the Army as a whole already had some experience. Ciphers had been used in the Civil War, and the War Department had employed a telegraph code since 1885. More recently, in 1916, the Signal Corps' Capt. Parker Hitt had published a manual on military cryptography However, it was Van Deman who made cryptology an adjunct of Military Intelligence. One of his first acts after setting up the Military Intelligence Section was to secure a commission for a youngish State Department code clerk, Herbert O. Yardley, so Yardley could head up a Cipher Bureau for Van Deman's organization. Although Hitt and a few other Regular Army officers possessed the necessary expertise in this field, all of them were needed for other duties when war broke out.
The Cipher Bureau, later redesignated MI-8, soon found itself caught up in a multiplicity of projects. Van Deman quickly saw that the Army had no means
30
of secure communications. The War Department Telegraph Code of 1915 in use for Army administrative communications was a cumbrous work designed to save telegraph costs rather than to provide security, and in any case, it was probable that the Germans already possessed a copy of the code book. Yardley's bureau was hastily directed to devise new enciphering tables for the code that once more made it practical for secret messages. MI-8 then prepared a completely new code that Military Intelligence as well as the rest of the Army could use. Regrettably, this was compromised as soon as it was issued, and a new code could not be prepared before Armistice Day However, MI-8's own communications system, which made use of the new enciphering tables, remained secure and was used throughout the war as a channel to transmit messages from the secretary of war and the chief of staff to the field.
MI-8 was to have more success in breaking codes than in making them. It soon began to attack agent communications. This led Yardley's unit into the arcane world of secret inks, regarding which the British provided help and advice. It was forced also to learn to read the diverse systems of shorthand employed in the United States and abroad. 17
MI-8's increasing responsibilities, together with the AEFs anticipated requirements in the area of codes, soon created a demand for trained cryptologic personnel. To meet the need, the Army turned to the only organization in the United States with cryptanalytic expertise. This was Riverbank Laboratories, a private research foundation set up by the eccentric philanthropist George Fabyan in Geneva, Illinois.18 One of Fabyan's hobbies was attempting to prove that Shakespeare's plays contained hidden cipher messages revealing that the works had actually been written by Sir Francis Bacon. Fabyan was wrong-no such cipher existed-but his obsession had led to the creation of a center for cryptanalysis at Riverbank. This organization trained three classes of Army cryptanalysts in late 1917 and early 1918. The arrangement ended only when MI-8 developed its own training program. Ultimately, two Riverbank instructors accepted Army commissions and went to France as cryptanalysts with the AEF
The cryptanalytic side of its work soon led MI-8 into other fields. To supplement physical interception of German messages, it established a radio intelligence service using selected Signal Corps personnel. These specialists monitored German diplomatic and agent communications, initially employing a chain of fourteen "radio tractors" strung out along the Mexican border and later using fixed intercept sites in the same locations. The radio intelligence service also had a large fixed station at Houlton, Maine, that monitored transatlantic German diplomatic communications.
31
The Military Intelligence Division
In April 1918 General Peyton C. March was recalled from France to become the new Army chief of staff. In June he ordered Van Deman, now a colonel, to go to Europe to inspect the intelligence operations of the AEF Van Deman left behind a functioning Military Intelligence organization that had a strength of 170 officers and hundreds of enlisted agents and civilians and was still growing. The expansion forced the War Department to again move its intelligence headquarters, this time to the Hooe Building in Washington, D.C. However, the full development of the wartime Army Intelligence organization was not yet complete. Military Intelligence reached its final organizational development in World War I as a result of March's experiences with Pershing's AEF
In August 1918 March restructured the General Staff, in the process raising the intelligence function to the status it enjoyed in France. He established the Military Intelligence Division (MID) as one of the four principal divisions of the War Department General Staff. The new division's enhanced prestige and responsibilities meant that Van Deman's successor, a Field Artillery lieutenant colonel from the AEF with the glorious name of Marlborough Churchill, was advanced to the rank of brigadier general. With these changes Military Intelligence had finally reached the position of institutional equality on the Army Staff that Van Deman had long advocated. In turn, this elevation in status permitted a more elaborate form of organization, with Positive and Negative Branches now controlling the various numbered sections.
One of the motives for the establishment of MID was March's desire to create an organization that could bring Army Intelligence training in the continental United States into line with the needs of the AEF Military Intelligence had heretofore neglected this area for a number of reasons, including its limited charter, its concentration on counterintelligence activities, and its lack of jurisdiction over training. The perceived deficiencies of training in the United States had previously led Pershing to demand that all commanders and staff officers, including intelligence officers, spend a lengthy period in France prior to the arrival of their divisions. To help rectify this situation, MID recalled an experienced intelligence officer from France to head a new Field Intelligence Section, MI-9, that would provide MID input to the training camps. However, the war ended before the plans devised by the new element could be put into effect.
In the meantime, MID's counterintelligence operations underwent a further expansion. The Negative Branch extended the scope of its operations across the seas in September 1918, taking over supervision of the counterintelligence operations of military attaches. It also created a new section to handle passport control duties previously shared jointly by MI-3 and MI-4. The State Department had primary jurisdiction over foreign travel by American citizens, but MID also wanted to screen all individuals traveling to Europe. Military Intelligence was particularly interested in checking the backgrounds of the large
32
numbers of welfare workers sent overseas to support the AEF by the Red Cross, the YMCA, the Salvation Army, and other organizations.
The same month, MID also gained control over military censorship within the continental United States, setting up another new section, MI-10, to handle the assignment. Originally Army censorship had been regarded as an independent function under a chief military censor, partially out of deference to the American tradition of freedom of speech. However, the general Army reorganization of August 1918 had made this an intelligence responsibility. MI-10's activities soon encompassed a wide variety of fields-in addition to censoring military mail and Army photographs, the organization also supervised a system of voluntary press censorship, ran a newspaper clipping bureau, accredited newspaper correspondents, monitored telephone and telegraph lines running into neutral Mexico, assumed direction of the radio intelligence service, and maintained liaison with other government departments. All this demanded a heavy commitment of personnel, and by the end of the war 300 people were on duty with MI-10.
In two areas the Negative Branch ventured completely beyond the normally defined boundaries of intelligence activity. The branch became involved in investigating graft and fraud within the Army and in finding ways to enhance military morale. in both cases, the transfer of these functions to intelligence came by default. Although the quartermaster general was originally assigned to deal with cases of graft and fraud, he had no trained investigative personnel. As a result, MI-13 was organized in August 1918 to assume the mission. The problem of Army morale was added to the MID agenda when Corps of Intelligence Police investigators included instances of low morale in their intelligence reports. Since MID had discovered the problem, it seemed appropriate to some members of the Army Staff that the division find a way to counter it. In this case, however, MID was able to reassign the function to a separate Military Morale Section of the General Staff in November 1918.
The armistice finally ended the headlong and rather undisciplined expansion of MID's activities. By this time, 282 officers, 250 CIP agents, and 1,100 civilians were on its staff. Van Deman's labors had succeeded in building up a massive intelligence organization that doubtless surpassed his wildest expectations. The establishment of MID had restored the intelligence function to the level of the General Staff, where it would remain from then on. However, in the process it had acquired broad ancillary functions the precise boundaries of which had yet to be defined rationally.
The American Expeditionary Forces in France
The American Expeditionary Forces in France built up an intelligence organization parallel, but not completely similar, to MID. When Pershing took command of the AEF, he designated Maj. Dennis E. Nolan as his intelligence officer.
33
Once in Europe, Pershing decided to adopt the French staff system throughout the AEF Intelligence became the second section, or "G-2," of Pershing's headquarters staff. Intelligence thus achieved a position of equality with other functional areas in the AEF a year before it would do so at the General Staff level in the continental United States, and Nolan ultimately became a brigadier general.
Pershing arrived in France in June 1917 with a small headquarters and a hastily formed division of infantry. By November 1918 the AEF had grown to a force of twenty-nine combat divisions and had opened its own front against the armies of Imperial Germany. The expansion and elaboration of the Army intelligence structure ran hand in hand with the growth of the AEF it supported.19 Initially the AEF heavily depended on help from the British and French in this field as in many others. Intelligence officers assigned to document exploitation and prisoner-ofwar interrogation were trained at the British intelligence school at Harrow until August 1918, when the AEF finally opened its own intelligence school at Langres, France. Students at Langres were provided with the unique opportunity to interrogate real prisoners of war as part of their training. From the start, the AEF trained its own photo interpreters and intercept operators.
By the end of the war, G-2 at Pershing's Chaumont headquarters had grown to a full-fledged theater intelligence center, engaging in a span of activities that was even broader than that of the Military Intelligence Division in Washington, since it also supervised deception operations and actively managed a propaganda campaign. In contrast, the work of MID's "psychological" subsection concentrated on coordination and training. The intelligence operations of Pershing's G-2 staff also overlapped and duplicated those of MID, because it produced its own political and economic intelligence as well as dealing in combat intelligence. Pershing considered that as a theater commander, his legitimate intelligence interests extended beyond the immediate Western front to cover developments on the Eastern, Macedonian, and Italian fronts. In Pershing's mind, any overlap with the activities of MID in the continental United States was justified because MID was an ocean away from the main battle and the transatlantic cable system had limited capacities.
The capstone of the intelligence pyramid, the Military Information Division, or G-2-A, produced finished intelligence reports and studies from the mass of information available from the AEFs tactical units and the other divisions of G-2. 20 The division was able to draw upon the full range of intelligence disciplines (human, photographic, and signals) to supply combat intelligence, and it also produced political and economic intelligence, mostly from open sources. The production activities of G-2-A had the result of involving the organization
34
in areas outside the field of pure intelligence: by the summer of 1918 G-2-A was also releasing the Army's daily public affairs communique and furnishing the War Department with a daily informational cable.
The base of this intelligence pyramid within the AEF began at the level of the battalion, the smallest unit with an intelligence staff officer, the S-2, and dedicated intelligence collection personnel.21 In addition to its S-2, each infantry battalion had a reconnaissance element consisting initially of 1 officer and 28 enlisted men, including 15 scouts, 11 observation post personnel, and 2 snipers. At the next level of the command structure was a regimental intelligence officer, with 8 additional observers at his disposal. Brigades were not authorized intelligence officers under the original scheme of organization, but in practice intelligence personnel were frequently detailed to these commands.
At the division level, there was a small intelligence section headed by the divisional G-2, who was assisted by a deputy for combat intelligence, a commissioned interpreter, a topographic officer, and various enlisted personnel. This was the initial level for interrogating prisoners of war and collecting enemy documents. Ground observation at this echelon was supplemented by spotting reports from Army Air Service balloons forwarded through division artillery channels. Under World War I conditions, divisions were responsible for keeping watch over the section of enemy front opposite them for a depth of two miles. In practice, however, the length and breadth of a division's span of interest was often much larger.
Balloons were used primarily for observing artillery fire, but balloon units also provided intelligence, especially under the static conditions of trench warfare that prevailed on the Western front.22 The main disadvantages of balloons were their vulnerability and their requirement for support personnel. A balloon in combat had an estimated life expectancy of fifteen minutes; although fliers did not wear parachutes in World War 1, balloonists did. It took a full company of 178 men to service a single balloon, because it took many hands to deploy a balloon in a breeze and because each company was also responsible for manning six antiaircraft machine guns to defend its fragile aerial asset.
The intelligence officer at corps level used a wider array of resources that allowed him to take responsibility for surveillance of the area between two and five miles beyond the enemy's forward line of troops. In addition to observation posts and balloons, he used aero squadrons equipped for visual and photo
35
graphic reconnaissance and sometimes even flash- and sound-ranging troops who detected and targeted enemy batteries. These were under the command of the corps artillery headquarters but under the staff supervision of G-2. Five of the seven corps that the AEF committed to combat were supported by such troops.23 The corps also had its own dedicated counterintelligence element consisting of twelve Corps of Intelligence Police sergeants.
The most sophisticated intelligence resources were concentrated in the field armies formed in mid-1918 and at GHQ, Chaumont. Each of the two field armies in the AEF had additional aerial reconnaissance units, including some that could operate at night.24 A topographic battalion allowed the army-level intelligence staff to draw up large-scale war maps, termed plans directeurs in the jargon of the day. Each Army Intelligence staff also contained a radio intelligence section that translated intercepted enemy messages. A larger radio intelligence element at GHQ engaged in cryptanalysis and supplied the subordinate Army sections with the necessary keying material to decode the messages. Intercept was provided by Signal Corps personnel who operated direction-finding and intercept equipment and manned listening posts directed against lowlevel enemy telephone and ground telegraph communications, a task facilitated by the use of induction coils rather than by direct wiretaps.25
Pershing's radio intelligence organization not only monitored enemy ground communications, but also could track the movements of enemy spotter aircraft through their transmissions. Searchlight platoons of the 56th Engineers attached to the AEFs Antiaircraft Service provided early warning of night air attacks. These units were equipped with multihorn and parabolic acoustical detectors that alerted crews to the approach of aircraft before they were in visual range.26
At the top of this layered intelligence structure was G-2-A at Chaumont, which processed and analyzed the information sent up from lower levels. Its efforts were supplemented by those of three other, more specialized, divisions of G-2. The "secret service," G-2-B, supervised both undercover collection operations and counterintelligence. Although most of the intelligence obtained by Pershing's headquarters about political and economic develop-
36
ments in Germany came from open sources, G-2-B did set up "information centers" in Switzerland, Denmark, and Holland that ran agent nets behind the enemy lines. The most valuable contribution of clandestine intelligence probably came from the reports of the "trainwatchers" who monitored rail movements of the German army.27 G-2-B's information centers competed directly with the MID-directed activities of the military attache offices in the same neutral countries and depended heavily on British and French assistance. The British and the French were also called upon for assistance in counterintelligence. However, the development of the Corps of Intelligence Police gave G-2-B an instrument of its own in this field. By Armistice Day some 450 sergeant investigators were on duty with the AEF, not only supporting the rear communications zone, but also providing intelligence coverage to corps and divisions at the fighting front.
Mapping was normally a function of the Corps of Engineers. But under combat conditions in France, it fell under the supervision of the intelligence staff, since battle maps included information about enemy as well as friendly forces. G-2-C, the Topographic, Map Supply, and Sound- and Flash-Ranging Division of G-2, essentially served to coordinate the activities of the 29th Engineers, a bizarrely structured regiment without a headquarters that provided the AEF with both topographic and sound- and flash-ranging personnel. One battalion of the 29th manned the AEFs large map-printing facility at Langres; another supplied topographic troops to the field armies; ranging companies from two additional battalions supported the targeting needs of the field artillery while providing collateral intelligence.
Finally, a fourth division, G-2-D, handled press and censorship matters. Censorship operations were an intelligence function within the AEF from the very beginning. While mail was censored at the unit level, a base censor's office in Paris conducted spot checks of regimental mail, operated a secret-ink laboratory, and censored mail written in foreign languages. Internally, the section was able to translate mail written in forty-nine languages, requesting outside help only for messages written in Chinese or Japanese. One additional function of this office was to censor letters containing information that servicemen did not want their immediate superiors to see. Such correspondence was mailed in specially issued blue envelopes.
Managing press relations was the division's second principal function. This included supervising the accreditation of war correspondents, censoring their dispatches, and making arrangements for their transportation and billeting. In view of practices in later wars, correspondents were held on a fairly tight rein.
37
Censorship principles insisted on accuracy; forbade releasing military information useful to the enemy; and prohibited news stories which would "injure morale in our forces here, or at home, or among our Allies" or "embarrass the United States or her Allies in neutral countries."28 On the other hand, correspondents had free access to the troops, and G-2-D refrained from using them for propaganda.
In addition, G-2-D carried out a wide spectrum of other information-related activities. It was charged with preparing propaganda to undermine German morale-by the fall of 1918 a rain of 3 million propaganda leaflets from G-2-D was blanketing German lines, distributed by plane and balloon and even by rifle grenade and patrol. And while it sought to undercut German morale, the division tried to preserve that of the AEF The famous troop newspaper, Stars and Stripes, was published in Paris under G-2 auspices. Finally, straying even further from any conceivable intelligence functions, the division supervised the Army art program, employing the services of eight soldier-painters, and was tasked with taking photographs for historical purposes and verifying the accuracy of their captions.
Communications Security in the AEF
Communications security within the American Expeditionary Forces in France was primarily a Signal Corps responsibility. Although MI-8 was responsible for preparing the War Department's codes, in France this task was assigned in December 1917 to the Signal Corps' Code Compilation Section, a small group located at the GHQ at Chaumont. This organization found itself confronted by an immense task. The Army entered World War I with few effective arrangements for secure tactical communications. The insecure and cumbersome War Department Telegraph Code was intended for administrative use, not battlefield communications. The Signal Corps' existing cipher disk was a simple celluloid device designed on principles as uncomplicated as those of the toy code-rings that once appeared in cereal boxes. When used with a running key, however, at first it seemed to offer unbreakable security. However, cryptanalysts at Riverbank and at MI-8 quickly discovered that messages encrypted this way could be broken faster than they could be enciphered. The British introduced the U.S. Army to the Playfair cipher, but this too was easy prey for cryptanalysts.
As a result, the AEF decided to devise a completely new system. The first efforts were unsuccessful. The Code Compilation Section produced a one-part trench code and a set of enciphering tables, but these proved to be impracticable for use in combat situations. What was needed was a method of encryption that would place as little burden as possible on communicators operating under
38
battlefield conditions.29 The solution finally adopted was the creation of a set of two-part codes with separate tables for encoding and decoding. In June 1918 the section produced the "Potomac Code," the first in a series of codes named after major American rivers. The "river" codes were issued to the First Army when it became active in August 1918. When the Second Army took the field in October, a separate "lake" series of codes was introduced to meet its needs. These codes, distributed down to regimental level through intelligence channels, carried the main burden of Army traffic, supplemented as necessary by a small number of more specialized codes. To ensure security, individual code books were replaced about every two weeks or upon evidence that they had been compromised.
Regrettably, the sophistication of the AEF's cryptographic systems did not equate with good communications security. The mere fact that codes were available did not mean that they were used. Communicators unfamiliar with codes showed a marked disinclination to employ them, and tactical units evolved various private, unsanctioned codes of their own. Moreover, the AEF relied on the telephone rather than the radio for the bulk of its communications, and officers repeatedly sent plain-language messages on tactical matters over unsecured telephone lines. To eliminate such problems, the AEFs general headquarters issued security guidelines for Army communications and established its own Security Service. Signal Corps personnel assigned to the Security Service monitored radio and telephone communications and reported violations of established procedures to control officers at the radio intelligence sections within the headquarters of the field armies. The G-2 then reported security breaches to commanders, but little could actually be done to punish offenders.30
Sideshows
The bulk of American forces committed overseas in World War I went to France. However, it is sometimes forgotten that there were some peripheral ventures. An infantry regiment and support troops were hastily deployed to the Italian front to help shore up the morale of the flagging Italian Army after its disastrous defeat at Caporetto in the fall of 1917. More importantly, two separate American Expeditionary Forces, miniatures of the larger AEF in France, were sent respectively to Murmansk and Siberia in the summer of 1918 as a result of the Bolshevik Revolution, which had forced Russia out of the war and
39
created conditions of chaos in Eastern Europe. Both expeditions were launched on ill-defined missions and given little background intelligence.
The reinforced regiment ordered to North Russia, originally assigned the limited mission of keeping a large store of war supplies from German hands, became involved in operations against the Bolsheviks as part of a much larger Allied effort and only evolved a G-2 section of its own shortly before the force was withdrawn in early 1919. However, Bolshevik propaganda and war-weariness after the armistice did confront the unit with serious problems of subversion.
A larger force under Maj. Gen. William S. Graves went to Siberia to help rescue former Czechoslovakian prisoners of war from the Bolsheviks and incidentally to keep an eye on the Japanese intervention force sent on a similar mission. Since the initial American contingent deployed from the Philippines and included elements from the Philippine Department's Military Intelligence Division, it incorporated an intelligence section from the beginning. Additional intelligence personnel from the continental United States joined the force later. Unfortunately, intelligence efforts were largely stultified by a conflict between General Graves and his G-2 concerning the nature of the latter's responsibilities. Graves pursued a policy of strict neutrality towards the Bolsheviks and vetoed any attempt to gather intelligence by using the services of the anti-Bolshevik "White" forces.31
Conclusion
World War I was the watershed in the evolution of U.S. Army Intelligence. Both in the War Department and in the field intelligence work was revitalized and placed on a footing of organizational equality with other major functions. The Army ventured into new fields of counterintelligence and cryptology and made use of the full spectrum of intelligence sources. Although some of thesesuch as prisoner of war interrogation, captured document exploitation, and ground reconnaissance-were traditional, the newer disciplines of signals intelligence, aerial photography, and collection through sensing devices were not. At both the War Department and theater levels, the definition of Military Intelligence was enlarged to include the collection of political, economic, and social data. Finally, intelligence activities were expanded to include the allied fields of deception and propaganda.
The professional intelligence field was still, however, in its infancy. Intelligence was still considered essentially a staff-level activity within the Army. No intelligence units as such were fielded, although the Army did deploy a topographic engineer regiment and aerial observation groups to France. A table of
40
organization for a field-army radio section-a communications intelligence unitwas drawn up, but never implemented, as Pershing chose to keep his limited intercept assets under direct GHQ control. Moreover, Army Intelligence was not yet considered an official career field. After the war, both Van Deman and Nolan became major generals, but they achieved their stars as tactical commanders, not as intelligence officers. Still, compared with the past, much had been accomplished. For Army Intelligence, World War I represented a great leap forward.
As an Army Intelligence organization was put together from nothing in the space of seventeen hectic months, many mistakes were made. MID largely concentrated its efforts on counterintelligence operations directed at a threat that proved to be largely imaginary.32 Yet its concern, however misplaced, reflected widespread perception among American leaders. The year before war broke out, even President Wilson's closest confidant, Colonel House, fretted that "there are more German reservists here [in the United States] than I thought."33 More telling is the criticism that MID spread itself too thinly by venturing into areas like fraud investigation and morale enhancement. Even some of its foreign intelligence efforts could be questioned. MID maintained files on areas as remote from the war effort as the Antarctic islands and on topics as broad as "Christendom." 34 Similarly G-2 in the AEF became deeply involved in press relations, a precedent which confused the duties of intelligence with those of public information in a way that would have an adverse application for Army Intelligence in the twenty years of peacetime following the armistice.
Nevertheless, the Army had at last established a permanent structure for meeting its intelligence needs. After 1918 the evolution of Army Intelligence would follow a twisted road, but it would never return to the marginal position it had occupied prior to World War I.
1 Powe, The Emergence of the War Department Intelligence Agency, pp. 82-84.
2 Finnegan, Against the Specter of a Dragon, p. 45.
3 Sir Basil Liddell Hart, Strategy (New York: New American Library, 1974), p. 6.
4 In his memoirs, Van Deman highlights his own role in bringing about the rebirth of Army Intelligence. The Final Memoranda, pp. 21-23. Marc Powe finds substantial continuities between the old Military Information Division and Van Deman's organization. The Emergence of the War Department Intelligence Agency, pp. 77-81, 102-03.
5 History of the Counterintelligence Corps, vol. 3, U.S. Army Intelligence Center, Fort Holabird, Md., 1959, p. 1.
6 Ibid., p. 4.
7 It should be noted that Van Deman's first intelligence assignment, in the Philippines, had been as a counterintelligence officer fighting insurrectos. There, he had "synthesized reports, analyzed captured documents, and provided pictures and descriptions of known revolutionaries." Brian McAllister Lynn, The U.S. Army and Counterinsurgency in the Philippine War, 1899-1902 (Charlotte: University of North Carolina Press, 1989), p. 155.
8 Intelligence officers professed to fear that "unless we proceed on an extensive and thorough scale, the enemy, with his existing system in the United States, will be stronger than we are right in our own army and we will be helpless." War Department General Staff, War College Division (WCD), MI-3, Provisional Counter Espionage Instructions, WCD 10148-37, INSCOM archives, Fort Belvoir, Va., Feb 1918, p. 3.
9 Ibid.
10 History of the Counter Intelligence Corps, vol. 3, p. 3.
11 As was later pointed out, the title of the organization was a little anomalous. "It was a `Corps' that was not a `Corps.' It was a `counterintelligence' organization but called an `Intelligence' organization and it was called `Police' when it had no interest in crime, as such, and no police power." History of the Counter Intelligence Corps, vol. 3, p. 114.
12 The Final Memoranda, p. 37.
13 The commander of the Corps of Intelligence Police in France was Lt. Royden Williamson. His recollections, "As It Was in the Beginning with the Corps of Intelligence Police," 11 Aug 53, are in RG 319, NARA.
14 Lt Col Frank Moorman, "Lecture Delivered to the Officers of the Military Intelligence Division, General Staff-Concluding Remarks by Brigadier General Churchill," 13 Feb 20, p. 16, Army Cryptologic Records, INSCOM.
15 Bruce W. Bidwell, History of the Military Intelligence Division, Department of the Army General Staff, 1775-1941 (Frederick, Md.: University Publications of America, 1986), p. 123.
16 War Department General Staff, Military Intelligence Division, The Functions of the Military Intelligence Division, General Staff (Washington, D.C.: Government Printing Office, 1918), p. 18.
17 David Kahn, The Codebreakers: The Story of Secret Writing (New York: MacMillan Co., 1967), pp. 352-54.
18 Special Research History (SRH) 29, A Brief History of the Signal Intelligence Service, p. 1, RG 457, NARA (hereafter, these documents will be cited by SRH number).
19 There has never been a scholarly study of intelligence work in the AEF. British Military Intelligence in World War I has been better served. See Michael Occleshaw, Armour Against Fate: British Military Intelligence in the First World War (London: Columbus Books, 1990).
20 Reminiscences of the order of battle specialist on Pershing's G-2 staff are contained in Samuel T. Hubbard, Memoirs of a Staff Officer (Tuckahoe, N.Y.: Cardinal Associates, Inc., 1959).
21 United States Army in the World War, 1917-1919, vol. 13, Reports of the Commander in Chief AEF, Staff Sections and Services (Washington, D.C.: Government Printing Office, 1990), pp. 1-10; Maj Gen Dennis E. Nolan, Dictation of March 2, 1935: Echelons of intelligence from the Front Line Back to G.H.Q., Nolan Ms, U.S. Army Military History Institute; Intelligence Regulations, American Expeditionary Forces, Aug 1918.
22 Balloons could ascend to an altitude of 4,500 feet, allowing observers to watch enemy positions eight miles away. Richard P. Weinert, A History of Army Aviation, 1950-1962: Phase I: 1950-1954 (Fort Monroe, Va.: U.S. Army Continental Army Command, 1971), p. 2.
23 A participant documented this specialized technical activity in Edward A. Trueblood, Observations of an American Soldier During His Service with the AEF in France in the Flash Ranging Service (Sacramento, Calif.: News Publishing, 1919).
24 The AEF experience produced the first clash between ground commanders and airmen over the priority to be given aerial reconnaissance. It would not be the last. Interestingly enough, the recalcitrant aviator was Col. William Mitchell. See Maj Gen Dennis E. Nolan, Dictation 38, p. 17 (1-31-36), Nolan Ms, U.S. Army Military History Institute.
25 War Department, Office of the Chief Signal Officer, Final Report of the Radio Intelligence Section, General Staff, General Headquarters, American Expeditionary Forces (Washington, D.C.: Office of the Chief Signal Officer, 1935). An anecdotal treatment of these activities can be found in E. Alexander Powell, The Army Behind the Army (New York: Charles Scribner's Sons, 1919), pp. 16-22.
26 William Barclay Parson, The American Engineers in France (New York: D. Appleton and Co., 1920), pp. 256-61.
27 Bidwell, History of the Military intelligence Division, p. 139. General Nolan felt that no more than 15 percent of his intelligence requirements were filled by secret agents, but felt that reports by the network of trainwatchers tracking German rail movements across the bridges over the Rhine were most useful. Memoirs, ch. 4, p. 9, Dennis Nolan Ms, U.S. Army Military History Institute.
28 United States Army in the World War, vol. 13, p. 86.
29 "In view of the fact that code work is frequently done under heavy bombardment and gas or in the critical moments of an advance, it does not seem advisable to add any additional burdens to code operators." War Department, Office of the Chief Signal Officer, Report of the Code Compilation Section, American Expeditionary Forces, December 1917-November 1918 (Washington, D.C.: Office of the Chief Signal Officer, 1935), p. 1.
30 James L. Gilbert, "U.S. Army COMSEC in World War I," Military Intelligence 14 (January 1988): 19-21.
31 The prominent historical novelist Kenneth Roberts served on the intelligence staff of the Graves Expedition. His recollections of the experience can be found in his autobiography, I Wanted to Write (Garden City, N.Y.: Doubleday and Co., Inc., 1949), pp. 98-112. I am indebted to Col. William Strobridge, U.S. Army (Ret.), for calling my attention to this fact.
32 There were also questions as to the efficacy of the whole effort. On the eve of America's entry into World War II, the officer who then headed the Corps of Intelligence Police provided a scathing evaluation of the World War I counterintelligence program. "In the United States the Corps of Intelligence Police was composed of many well-meaning but inexperienced officers, enlisted men, and civilians and their unorganized efforts accomplished practically nothing. It is said the organization in the United States may have caught one spy." History of the Counter intelligence Corps, vol. 4, p. 122.
33 Finnegan, Against the Specter of a Dragon, p. 149.
34 Memo for Acting Director, Military Intelligence Division, 24 Apr 20, MID Documents.
page created 10 September 2001