The prosecution of Captain Alfred Dreyfus for espionage in France in 1894 remains one of the most instructive cases in legal history. Dreyfus was first presumed guilty; his fluency in German, exceptional memory, and broad knowledge were then turned into evidence against him. Today, similar concerns surround the investigation into the death of Muhsin Yazıcıoğlu: investigators are accused of searching for connections that support a predetermined conclusion. An investigation that works backward from a chosen suspect, rather than following the evidence to whoever is responsible, moves us further away from justice.
Can speaking German fluently be evidence of a crime? What about having an excellent memory? Or being exceptionally well-read? In France, all these qualities were once used as evidence of treason and espionage.
In 1894, a note containing information about the French army was discovered in the wastebasket of the German military attaché in Paris. Investigators began searching for a spy, and suspicion soon fell on Captain Alfred Dreyfus.
What followed became one of the most instructive episodes in legal history. First, Dreyfus was judged guilty. Then, almost everything about his life was turned into evidence supporting that judgment.
The first expert assigned to examine the document said its handwriting bore no resemblance to Dreyfus’s. In a normal investigation, this would have been considered a significant finding in the suspect’s favor. Not this time. When the expert failed to deliver the desired conclusion, other experts were found and reports were commissioned to support it.
Even that was not enough. To strengthen the case, a “secret dossier” was submitted to the court and withheld from the defense. Years later, one of its documents would be exposed as a forgery, while another proved to be correspondence with no demonstrable connection to Dreyfus.
Then came Dreyfus’s personal qualities. His “remarkably powerful memory,” his “unusually broad knowledge,” and his fluency in German were all treated as signs that he was capable of espionage.
Notice the distinction: Dreyfus was not identified as a spy because he spoke German. His knowledge of German became evidence of a crime because he had already been presumed to be a spy.
Decide the Outcome First, and Everything Becomes Evidence
Imagine that you begin by deciding: “This man is a spy.” From that point on, you can interpret anything you find to fit that conclusion.
He speaks German: So he can communicate with the Germans!
He has an excellent memory: So he can retain classified information without writing it down.
He is exceptionally well-read: So he has the intellectual capacity to conduct a complex espionage operation.
His handwriting matches: There is your evidence. His handwriting does not match: He must have changed it to conceal his identity.
Within this way of thinking, no finding can establish innocence. Everything can be made to support the same conclusion. Even the absence of evidence can become evidence: “Look how expertly he covered his tracks.”
Can this be called an investigation? No. It is the manufacture of justifications for a verdict already reached.
The Real Suspect Emerges
After Dreyfus’s conviction, Lieutenant Colonel Georges Picquart took charge of French military intelligence. While investigating a new document, Picquart realized that suspicion should instead fall on another officer: Major Ferdinand Walsin Esterhazy.
He investigated Esterhazy’s contacts with the German embassy and compared his handwriting with the document used to convict Dreyfus. More importantly, he discovered that some documents in the secret dossier against Dreyfus were forgeries.
He alerted his superiors. What happened next?
The system did not say, “We convicted the wrong man.” Quite the opposite: attempts were made to silence Picquart. He was sent to distant postings and threatened. Esterhazy, meanwhile, was tried and acquitted.
By then, the issue was no longer simply whether Dreyfus was guilty. Once a Jewish officer had been publicly branded a German spy, reversing course became difficult for both the government and the army. Exonerating Dreyfus would damage the reputation of the General Staff and jeopardize the war minister’s political future.
Protecting the institution’s version of the truth had become more important than discovering the truth itself.
Now Consider the Yazıcıoğlu Investigation
There is no question that the circumstances in which Muhsin Yazıcıoğlu and five others died on March 25, 2009, must be investigated fully. Serious questions have remained unanswered for years. Any negligence must be investigated. Any suppression or destruction of evidence must be investigated. Any sabotage must be investigated. If an organization planned their murder, that plot must be exposed in full.
But there is only one way to do this: follow the evidence to those responsible. Do not begin with a chosen culprit and work backward.
In a statement about the investigation in July, Justice Minister Akın Gürlek said the case involved “clandestine organizational structures,” “organizational ties,” “a connection to Adil Öksüz,” and “ByLock correspondence.” He also said that many suspects had previously been investigated or faced proceedings in connection with what he called “FETÖ,” the Turkish government’s designation for the Gülen movement.
He subsequently went further. He explicitly described the incident as “murder,” placed Hrant Dink’s murder in 2007 and Yazıcıoğlu’s death in 2009 within the same narrative of an organization’s development, and alleged “very deep connections” between them.
A “Connection” Is Not the Same as Evidence of Murder
Which of the connections cited by Minister Gürlek demonstrates that Muhsin Yazıcıoğlu was murdered and that the individuals named participated in a plot to kill him?
The most important lesson of the Dreyfus affair is this: once you decide the outcome in advance, even ordinary facts can become evidence of a crime. Dreyfus spoke German. Speaking German was not a crime. But if you begin by accepting that “Dreyfus is a German spy,” his knowledge of German suddenly takes on a different meaning.
The same method is being used to make accusations today. The fact that someone met with a person from a particular community is not, by itself, evidence of murder. But if you first declare, “The Gülen movement committed this murder,” every relationship with that community begins to acquire a different meaning.
A phone call starts to look incriminating. A family relationship starts to look incriminating. Having worked at the same institution starts to look incriminating. Even a person’s name appearing in an entirely unrelated investigation years later can be read as evidence of involvement in an earlier death.
A Simple Test
One of the easiest ways to expose a flawed method is to apply it to another group. Suppose we begin with this conclusion: “The government had Muhsin Yazıcıoğlu killed.”
I offer this assumption not as an allegation, but solely to illustrate how the method works.
Now let us begin investigating.
Who was in power at the time? The Justice and Development Party, or AKP. Who controlled the intelligence services? The government. Who was in charge of the police? The government. To whom did the public agencies responsible for search and rescue operations answer? The government. Under whose administration had the officials involved in the subsequent investigations been appointed? The AKP’s.
Next, let us pull the phone records, examine political relationships, and map the network of contacts. Finding “suspicious connections” will not be difficult.
Then let us draw a chart and label it “The Big Picture.”
Does any of this establish that “the government committed the murder”? Of course not. What we have established is association, not causation. Association does not prove a crime. Using this method, you can identify any culprit you choose.
Start with the Republican People’s Party, or CHP, if you wish. Or the Nationalist Movement Party, or MHP. Or another political group, an intelligence service, a foreign state. First decide the conclusion. Then sift through millions of phone calls, thousands of public officials, countless political relationships and social contacts, and events spanning many years. Select only what supports your conclusion.
In a sufficiently large pool of data, a connection can be found between almost anyone and anyone else. The question is not whether a connection exists, but what it proves.
The Real Lesson of the Dreyfus Affair
The injustice inflicted on Dreyfus went beyond his conviction on fabricated evidence. The deeper problem was that the investigation had become impossible to disprove. When evidence of innocence emerged, the system protected itself rather than taking that evidence seriously. When the real suspect appeared, Picquart—the man investigating him—was treated as the problem.
The best way to assess the integrity of an investigation is to ask what happens when evidence emerges that could undermine its case. Is that evidence included in the file, or kept out?
The same question must be asked of the Yazıcıoğlu investigation. If a finding contradicts the predetermined narrative, will investigators change course? Will they move from evidence to conclusion, or from conclusion to evidence?
The difference between these two approaches is actually very simple.
In the first, the question is: Why did Muhsin Yazıcıoğlu and those accompanying him die? Every possibility is investigated, and investigators follow whichever explanation the evidence supports. In the second, the question is different: How did the Gülen movement kill Muhsin Yazıcıoğlu? The moment you ask that question, you already have your conclusion. All that remains is to assemble the pieces.
A phone record. A ByLock user. An old investigation. An acquaintance. A relative. Suspicious timing. Put them all on the table, and call it “The Big Picture.”
If the Aim Is Truly to Solve the Case
Akın Gürlek says, “We have a solemn duty to solve this murder.” Such determination to uncover the truth behind a suspicious death is certainly commendable. But solving a case and attributing it to a predetermined culprit are not the same thing.
A narrative assembled from numerous relationships, contacts, and coincidences may lead to the accusation of innocent people rather than the resolution of the case. The harm does not end with those wrongly accused: it can also prevent the discovery of the actual perpetrators and the connections that truly matter. An investigation whose conclusion is settled in advance delivers justice to no one.
A similar method was used in the Hrant Dink murder case to construct a narrative of organizational involvement. Today, we see the same approach repeated in the Yazıcıoğlu investigation: the same methodological flaws reproduced with different names and different evidence.
Legal history is full of examples showing how dangerous it is to identify the culprit first and then interpret every fact to fit that conclusion. Yet repeating past mistakes as an investigative method seems far easier than learning from them.
If the genuine aim is to establish the truth about this death, the task is clear: investigate whom and what the evidence points to, without accepting any conclusion in advance, rather than searching for connections that confirm a predetermined culprit.
