I come from a corporate environment, and one of the biggest problem we have serving users is spending time looking for things that are not in the archive because they were never supposed to be there, but the user, who is not privy to those disposition decisions, never knew that.
In many corporate environments, that kind of decisioning is never written down. I can count on all my fingers and toes the business people I have spoken to who lack knowledge management systems and even simple governance guides, let alone business rules documentation. I once became the only operational employee of a startup that had never even printed an employee handbook or kept an archive of email to refer to the decisions the executive team had made using that channel. So I can tell you that outside of academic and GLAM+ environments, it is extremely hard to find documentation of disposition decisions, even when there is a DAM Librarian in-house or an Operations Director who is conscious and careful about the documentation of the decisions they make, and tries to write documentation for every new workflow or business rule.
Things can change fast in corporations, even in the archive. You'd think things like metadata schemas, taxonomies and ontologies would remain static...but they simply do not. One new executive vice-president can make it necessary to change everything. That is going to happen. And it needs to be traced and every relationship change tracked in a knowledge layer that is separate from assets but easily discoverable and searchable, because the questions will come up.
"Every archive preserves evidence of what was kept. The computational archive will also need to preserve evidence of what was intentionally forgotten." For the evidence of what was intentionally forgotten, what evidence is needed beyond the disposition authority and possibly the appraisal justification for that authority? Archives rarely have sufficient resources to preserve and provide access to their permanent holdings. What additional burdens are you suggesting they take on?
Why does the computational archive "need to preserve evidence of what was intentionally forgotten"? What knowledge do you envision the end user gaining from such evidence?
History has always been written based on whatever records have survived until the time the history is written, and not all of those records will be in the custody of archives. The historian rarely accesses all of the available evidence.
Mark, thank you. I think we're actually closer in our thinking than my wording may have suggested.
I'm not arguing that archives should preserve temporary records or take on the impossible burden of documenting every destroyed item. The disposition authority and appraisal rationale remain the primary evidence of intentional forgetting.
What I am suggesting is that, in a computational environment, those records of appraisal and disposition should be explicitly related to the archival corpus they shaped. Today, we often know what permanent records survived, and separately we know the schedules that authorized destruction. What is frequently missing is a machine-readable relationship between the two.
That relationship matters because computational systems increasingly infer meaning from what they can observe. If the machine sees only the surviving records, it can easily mistake the archival corpus for a complete representation of organizational activity. Knowing that entire classes of records once existed, why they existed, why they were judged temporary, and how they related to the permanent record provides important context for both AI systems and human researchers.
So the "evidence of what was intentionally forgotten" is not the forgotten records themselves. It is evidence that those records once occupied a defined place within the documentary universe and were intentionally removed according to governance decisions. In other words, the archive should preserve not only evidence of what an organization chose to remember, but also evidence of the choices that produced that memory.
That distinction becomes much more important when the primary reader is not a historian carefully weighing sources, but an AI system constructing relationships across millions of records.
I come from a corporate environment, and one of the biggest problem we have serving users is spending time looking for things that are not in the archive because they were never supposed to be there, but the user, who is not privy to those disposition decisions, never knew that.
In many corporate environments, that kind of decisioning is never written down. I can count on all my fingers and toes the business people I have spoken to who lack knowledge management systems and even simple governance guides, let alone business rules documentation. I once became the only operational employee of a startup that had never even printed an employee handbook or kept an archive of email to refer to the decisions the executive team had made using that channel. So I can tell you that outside of academic and GLAM+ environments, it is extremely hard to find documentation of disposition decisions, even when there is a DAM Librarian in-house or an Operations Director who is conscious and careful about the documentation of the decisions they make, and tries to write documentation for every new workflow or business rule.
Things can change fast in corporations, even in the archive. You'd think things like metadata schemas, taxonomies and ontologies would remain static...but they simply do not. One new executive vice-president can make it necessary to change everything. That is going to happen. And it needs to be traced and every relationship change tracked in a knowledge layer that is separate from assets but easily discoverable and searchable, because the questions will come up.
"Every archive preserves evidence of what was kept. The computational archive will also need to preserve evidence of what was intentionally forgotten." For the evidence of what was intentionally forgotten, what evidence is needed beyond the disposition authority and possibly the appraisal justification for that authority? Archives rarely have sufficient resources to preserve and provide access to their permanent holdings. What additional burdens are you suggesting they take on?
Why does the computational archive "need to preserve evidence of what was intentionally forgotten"? What knowledge do you envision the end user gaining from such evidence?
History has always been written based on whatever records have survived until the time the history is written, and not all of those records will be in the custody of archives. The historian rarely accesses all of the available evidence.
Mark, thank you. I think we're actually closer in our thinking than my wording may have suggested.
I'm not arguing that archives should preserve temporary records or take on the impossible burden of documenting every destroyed item. The disposition authority and appraisal rationale remain the primary evidence of intentional forgetting.
What I am suggesting is that, in a computational environment, those records of appraisal and disposition should be explicitly related to the archival corpus they shaped. Today, we often know what permanent records survived, and separately we know the schedules that authorized destruction. What is frequently missing is a machine-readable relationship between the two.
That relationship matters because computational systems increasingly infer meaning from what they can observe. If the machine sees only the surviving records, it can easily mistake the archival corpus for a complete representation of organizational activity. Knowing that entire classes of records once existed, why they existed, why they were judged temporary, and how they related to the permanent record provides important context for both AI systems and human researchers.
So the "evidence of what was intentionally forgotten" is not the forgotten records themselves. It is evidence that those records once occupied a defined place within the documentary universe and were intentionally removed according to governance decisions. In other words, the archive should preserve not only evidence of what an organization chose to remember, but also evidence of the choices that produced that memory.
That distinction becomes much more important when the primary reader is not a historian carefully weighing sources, but an AI system constructing relationships across millions of records.