2026 AskQC office hour member Q&A

12 May 2026

If a record is unmerged, is it at risk of being merged again?

If nothing changes in the record and it's unmerged and stays exactly as it was before it was unmerged, then it's very likely that it would be merged again. Because DDR has changed over the years, it is possible that a merge that happened years ago would not happen today, but for recent merges, it is recommended the unmerged record be updated with an edition statement or other data elements to prevent remerging.

Does OCLC proactively check for incorrect merges? If so, on what basis? Is there any way to proactively check for incorrect merges?

If we see a particular pattern to merges that are happening, we will try to investigate and see if we find other candidates that were improperly merged. But, considering the volume of merges that happen on a regular basis, it's really not easy for us to proactively check for incorrect merges. We definitely rely on our members to let us know if they see incorrect merges.

Is there a specific reason why 1830 was selected as the cut-off date for earlier items not being merged?

This date was chosen in consultation with the Rare Books and Manuscripts Section of the Association of College and Research Libraries. Some things that happened in the history of printing make that a good date to use for books. You'll notice that there's a different date cutoff for cartographic materials, and that was also done with consultation of members of the map cataloging community.

Does the software cover non-English cataloging (e.g., Spanish records), or just English?

DDR and the ML model cover all languages. Because they are evaluating the entire MARC environment, it is possible that if you're seeing multiple records for the same resource but differences in the cataloging make the software think they are not duplicates. But you can always report those to bibchange@oclc.org and we file them into our queue for manual merging as well.

Why are there excessive ISBNs on records? Is this a merge issue or manual cataloging?

There are a combination of factors causing this. Sometimes a cataloger adds an ISBN to a record incorrectly, and sometimes vendor-supplied records have ISBNs that belong to other resources. ISBNs are transferred from the merged record to the retained record (unless it is the same number). If a record that gets merged has an ISBN that doesn't belong to it, that will transfer to the record that is retained. So, merging is not creating the incorrect ISBNS–it is just transferring them in the merge. If you see an ISBN that doesn't seem to belong there, put it in subfield $z and report it to bibchange@oclc.org. Consult the BFAS 020 page for instructions on when to use subfields $a and $z in field 020. It is common for an online resource record to have the ISBN for the print resource in subfield $z and vice versa.

How does the system decide which record is retained in a merge?

The decision on what record is retained can depend on the fullness of a record, the number of fields present, how many holdings are present on the record. There are a lot of factors that go into it, so it's not very straightforward, but we try to evaluate the records and keep the most complete, beneficial record as the retained record.

Do libraries using WorldShare receive special cataloging training from OCLC?

OCLC provides free training for products, including WorldShare Record Manager. The main training page is here: https://help-it.oclc.org/Librarian_Toolbox/OCLC_training. Note that there is live training you need to register for and recordings you can view at any time. However, this is training on the products, not how to catalog. There is some freely available training available on the Library of Congress Catalogers Learning Workshop webpage.

When can we remove erroneous ISBNs ourselves vs. contacting bibchange?

It's very important to know what qualifies as an erroneous ISBN. Some ISBNs must stay on a record in subfield $z—for example, if a publisher reuses an ISBN and it appears on the item, it should not be removed. If it's clearly incorrect—for example, an ISBN on a record for a resource published in 1936—you would be justified in removing that. With full cataloging permission, you are allowed to edit field 020 in records, including PCC records. If you are ever unsure about whether the ISBN should be there or which subfield it should be in, email askqc@oclc.org for help.

If you have a record that has an ISBN that a publisher has erroneously added to the item, should we annotate this also in 500?

ISBNs appearing on a resource that belong to a different publication should be recorded in 020 subfield $z so users can still find the resource searching that ISBN. Optionally, you can also add a 500 field note to better explain why the ISBN is incorrect for that resource.

This might be too specific to our library but we did a reclamation project in 2023 and had all our LHRs removed as well. I get updated records from WorldShare for resources we don't own and records we don't have in our catalog. Why are holdings attached to records we don’t own and LHRs still on records when we had them removed?

We don't know why that is happening, so we recommend contacting the data analyst you worked with on your reclamation project. If you don't remember who that is, email us at askqc@oclc.org and we will get you in touch with the right person to help.

What is the best method for cataloging volumes with more than one distinct work?

There's not one specific method. A lot depends on how the title page is presented and cataloging choices at your institution. There are two basic approaches, and neither is wrong. One is more popular in North American cataloging: create one record for the entire volume and provide information about the contents. Titles may appear in the 245 or in a 505 field, and you can provide authorized access points for works in 7XXs. This is the approach described in original RDA 1.5.2 Comprehensive Description. Another method analytical description, where each title is cataloged as a part on a separate record. In original RDA this is described in 1.5.3 Analytical Description. I do not see this much in records from US and Canadian libraries, but I do see it in records from some European libraries. Assuming you go with the comprehensive description approach, what you record as the title proper depends on how the information appears on the preferred source of information–it depends on whether there is a collective title or if there is no collective title and titles for each work are listed on the title page. If it is the latter case, I recommend looking at the BFAS 245 page in the ISBD punctuation section for how to format your 245 field. Both of these methods assume you are asking about works published together. In rare book cataloging there is a situation called "bound-with" where resources separately published are later bound together. If you have something like that, please follow up with us at askqc@oclc.org and we can explain about how that works and provide some examples.

Do you have any guidance on cataloging early printing material where there are records that have local bibliographic data on the bibliographic record, for example, a Thomas Jefferson-owned copy?

We have instructions about this in BFAS 3.4, Local Information in Records. Sometimes copy-specific information should go in local bibliographic data (LBD) or a local holdings record (LHR) so it doesn't appear in the WorldCat records. These situations are situations where the information is primarily of interest to that library. For other situations like your Thomas Jefferson example, that kind of information can be recorded in the WorldCat records because it is important to historians, bibliographers, and others. When you add that information in a WorldCat records, you use the regular MARC field but add subfield $5 with your MARC organizational code to indicate it is copy-specific. It's never wrong to keep this kind of information as local data if you prefer not to have it displayed widely in WorldCat. You can contact askqc@oclc.org for guidance if you have a specific situation, you would like us to provide recommendations on.

21 May 2026

So many dupes for online books, and I'm wondering why?

Brief records may not get merged because they don't have enough data points for comparison. Also, records with incorrect coding, such as online resource not being coded as Form “o” (for online resource), can confuse the system. Just recently, with our machine learning de-duplication, we have adjusted some of the processing to try to adjust such records that are being referred to. So, what we are doing is allowing a little more variance in the records we merge, in certain categories, to allow for some of those challenges that DDR has in identifying these records as duplicates. We are hoping that that will help address some of the brief, minimal level, pre-publication records in WorldCat.

If having records that are too brief mean they can't be merged, why are these super brief records accepted at all?

We have criteria that a record has to meet in order to be added to WorldCat. So, it has to have enough elements and the description to be full enough to identify it. The criteria is a little bit elaborate and you can find the criteria through our data sync batch loading processes, Data Sync Processing, Sparse record criteria.

I know these are separate initiatives at OCLC, but are the algorithms for creating new OCLCKB brief records in sync with the merging algorithms, so that the new brief records aren't duplicating existing records for e-resources?

Those records are being based off of KBART data for use within the knowledge base. That's why those are brief, but they still meet the sparse record criteria. The metadata used to create those OCLC KB records is very different. It is KBART, so the information is not going to be as complete, so there's a lot of issues around that. When you have data that is not as complete, it's brief and it's harder to merge. But the KBART metadata doesn't make it difficult to match to other things that are in the database.

How does OCLC perform their quality control on merged records? Do you wait for a member to report something?

We do rely on our users to report merges when they look strange. Once you've reported a bad merge, we can look at what happened, whether it's something that needs to be adjusted in the software, especially as we're developing the machine learning model. That lets us expand from there to see if it's a pattern, if it was a one-off, or if it happened because of manual data that was entered incorrectly. But yes, we don't have any way of identifying it other than users letting us know.

Lately I am seeing a lot of very incomplete records with an author and title in ALL CAPS. Does someone come back later and update these records?

If you are seeing a record that doesn't have a lot of data elements and it is encoding level 3 with data in all caps, that's usually a record that was provided by a vendor. This is a balancing act about accepting these records. Unfortunately, there isn't a team of cataloging elves at OCLC that go back through the database and correct the capitalization of these. Libraries who catalog in WorldCat will upgrade the records and correct the capitalization when they catalog for their collection. While the capitalization in these records can seem annoying, the records provide a place to start for that resource and add holdings to.

Regarding 020s as a possible match point -- I've noticed that some of the more widely used records tend to collect 020s including those for other editions, why is that?

Some of that is a lack of understanding the work that the 020 field is doing. The only way it comes into a record is if it was present on some version of that resource, or someone's added it because they wanted to use that record for their item and they are not really thinking through the fact that the ISBN may not apply to it. That's where I've seen people go, a book that was published in 1932 and someone adds an ISBNs on the record that really shouldn't be there. And it can create issues with merging because DDR or the merge software might think about it and start seeing potential matches that shouldn't really take place. So I definitely encourage people, if you see ISBNs that don't make sense, put them in subfield $z and report them to us and we can look into what is going on.

How is OCLC planning on dealing with the upcoming changes to LCSH terms for Indigenous peoples? My understanding is that some of the changes will be splits that cannot be one-to-one changes.

We'll need to see what happens once the headings are changed and the results are published in the authority records. We've dealt with splits before, meaning that what was one authorized heading is now two headings, and sometimes it's appropriate for a record to have both of the new headings. Sometimes is appropriate for the record to have just one. So we're going to need to evaluate the situation when we see it and make some decisions. It's something we've handled before and will handle again and do our best to deal with that.

If a library adds $v to subject headings on an LC record, if LC changes their copy and redistributes it, will the $v's be lost?

That is a very complicated question to answer. The very short and unsatisfying answer is, “It depends.” There's a lot of issues to consider. One would be the encoding level of the record, the Library of Congress (LC) record. One would be, what else has happened to the record in between the time that LC adds it to WorldCat and when they redistribute it. So it's really not a question that we can answer easily. It's possible, yes. It's possible that the subfield $v might be lost. It's possible that it might be retained. We're evaluating this situation. We're taking a look at what's happening to the records regarding subfield $v and seeing what the impact is. It's important to remember that there's a lot of records that do not have form subdivisions in subfield $v at all and so there's no data loss there. But if there are some specific records you see where you think that there should be a form subdivision, you can add that if you have the capability with your cataloging authorization, or you can email bibchange@oclc.org to update the records if you are not authorized to do so.

The change from Indians to Indigenous peoples and Indigenous peoples—America is now live in LCSH. Is OCLC planning to wait for all related headings to be updated before making edits to the database?

No, we're not necessarily going to wait for all of those related headings to be changed in LCSH before we make updates to the database. The special LCSH list for September 2025 was processed last week. for the one-to-one headings, bibliographic records with controlled headings will be updated automatically. For the heading spit with "Indians" into "Indigenous peoples" and "Indigenous peoples $z America," we will need to evaluate bibliographic records and determine how to appropriately update them as some might need just one of the headings and some might need both. Whenever there is a heading split like this, we have to look at that situation and determine how to update the bibliographic records correctly. In the meantime, if you come across any records in you regular cataloging that you think should be updated, please feel free to do so.

14 April 2026

How are these codes used or usable in Connexion and Record Manager?

You add this information by creating a new MARC field in Connexion or Record Manager. Specifically, you add a 041 field, then enter the appropriate language code from the relevant code list using the correct subfields. After that, you include subfield $2 to identify the source of the code. For non-MARC language codes, you simply record them in a separate 041 field following this same pattern.

I noticed the codes were also used in the 650s. Could you speak about their use as well?

This is a little confusing because, in those 650s with second indicator 7, what they’re showing is not language codes being used as language codes. What they’re showing is a different vocabulary being used for subject headings. When you use 650 with second indicator 7, you can use a different subject vocabulary than LCSH or NLM subject headings. I’m not familiar with this subject vocabulary off the top of my head, but I’m assuming that is how that subject heading for that language is constructed in that vocabulary.

Is there any chance OCLC might automate some corrections to 546 fields?

We can take a look at that and see what’s in WorldCat to see what we can do. If you have other examples, feel free to send it to askqc@oclc.org.

Is it possible to search in Connexion to find records with specific $2 codes?

For indexing in WorldCat, I believe the subfield $2 codes are not indexed, but you can still, if you find a code from the list, search by the specific code from that code list.

If the second indicator is zero, there wouldn’t be a subfield $2, so you couldn’t search that. You could just search MARC language code values. If it’s second indicator 7, you can do a search on the code, like ISO 639-3, but we think there’s a little bug in it. It worked for Glottolog, but it might not work consistently. You may want to do something else besides just searching the language code, like a specific value.

Will the more granular languages from ISO 639-3 display or be indexed or faceted in WorldCat?

I believe they should all display in WorldCat Discovery, but I will check on that. We verified with Discovery staff who gave the following statement:

"No, they will not currently be faceted. The Language facet in WorldCat Discovery is driven by the standard MARC Code List for Languages (specifically looking at the 008/35-37 and standard 041 fields). The PCC guidelines explicitly acknowledge this system limitation on Page 1: "Many current systems can only handle MARC language codes or require a code in the language fixed field (008/35-37). Non-MARC language codes cannot be input in 008/35-37. In order to ensure that PCC records are usable in all systems, the guidelines that follow instruct PCC catalogers... to encode all languages using MARC language codes and, optionally, also using ISO 639-3 language codes." Because WorldCat's language facet relies on the MARC list, a highly granular ISO 639-3 code will not create a new, specific facet. For example, if you are cataloging a resource in "Fuliiru" (ISO 639-3 code flr), WorldCat will not generate a facet for "Fuliiru." Instead, it will look at your required MARC code (bnt for "Bantu (Other)") and group the item under the broader "Bantu" facet. Will they be indexed? Currently WorldCat Discovery still uses the older MARC language codes for language faceting/filtering, so even if a very specific ISO 639-3 language code is added (for example, Fuliiru), Discovery will usually group it under the broader MARC language category instead of showing a separate “Fuliiru” language facet. The detailed ISO 639-3 code is still stored in the record and may help with general discoverability, but it does not currently create its own dedicated language facet in Discovery."

23 April 2026

What would be the best practice for adding a language code to an international subject heading such as Homosaurus, which is published in English and Spanish now? Can you add a language code to a 650 field?

You can provide a data provenance subfield in field 650, that would be subfield 7. When you do that, you can specify the language and script of that heading with a BCP 47 code. That would be how you could specify a language of the subject heading. I have not seen that done, but it is technically possible.

For subjects that have a subfield 2, you can add a slash after the subfield 2 code and then have the MARC language code after that. If it’s from the Spanish version, you could add a slash and then the language code.

Where can you find a list of the codes that are valid for use in subfield $2 of a 041 field?

If you visit Language Code and Term Source Codes, it will bring you to a page where you can click on some of the source codes. The slides also have all the links to the websites on them so that you can go in and check out the list as well.

Will the MARC code list be updated with these new codes?

The MARC language code list is maintained by the Library of Congress, and they may add new codes, but that would be up to them to do. They are available to use separately in another 041 field if needed.

ISO 639-3 examples list Ethnologue, Glottolog, and Wikipedia—does that mean those are the codes used in those resources?

When you look up a code on the ISO 639-3 page and see those listed, that just links to those pages. Glottolog has its own codes.

I’ve noticed fewer FAST headings in WorldCat records lately. Are they still being added and used?

We’re working on adding them to any records that don’t have them.

Are you anticipating an increase in non-LC language codes in 041 fields given the recent MARC revisions?

We have seen an increase since the publication of the PCC guidelines for ISO 639-3. As we transition more into linked data, it is quite possible that we will get more of these other non-MARC language codes in our bibliographic records.

That proposal was approved but hasn’t been published in the MARC standard yet, so you won’t see any impact to WorldCat records for a while. But it may raise awareness of the availability of other language codes.

I have cataloged books in shorthand. The language is English, so I’ve used English in the fixed field, but is it appropriate to use 546 for a language note like 'Pitman shorthand'?

Field 546 is intended to be used as a textual explanation of the language or notation. If it helps bring out the specific type, it might be useful.

Is there a way to code for the shorthand script using one of the newer methods?

I don’t think shorthand is technically considered a script because it’s still Latin letters. I would have a 546 note that says “In English written in Pitman shorthand”, but I would not consider that a script.

When is it appropriate to use 546 text for English when it could be implied from the fixed field?

546 is intended to explain what’s going on with the language of the resource. The fixed field codes are meant to be used by the computer in the background. For more complex situations, it’s useful. You can still use 546 for English if you decide as a library you want to do that.

Is there a MARC field for ISO script codes, or do they go in 041?

You can combine language and script using RFC 5646 in the 041 field. You can also specify the language and script for the text in a particular field using data provenance, such as subfield 7.

You can also see script codes recorded in field 066. This is not done all the time, but you would primarily see it in non-Latin script records.

10 March 2026

Isn’t this just a work-around from dealing with NACO?

So, this is complimentary to NACO as it exists now. NACO establishes the authoritative string for a given person. The string is very helpful in gathering resources together as entities can be as well via the URI. Entity management is di􀆯erent in that it’s managing the actual person. We start to get into describing things about that person that can then be used to link to other people, works, and topics. NACO has its role and this is one way that you can manage entities along with something like Wikidata or ISNI, outside of that flow. This also works in situations where you might not have a controlled name. For uncontrolled names that don’t have NACO records, you could still add a subfield 1 to point to this entity, which provides another one of those aggregation points.

So “Description” is free text?

Yes, the description is a free text field and one of the fields that folks are interested in developing best practices for.

Is there a set of best practices/guidelines for handling things like titles? As a NACO participant, it’s just very di􀆯erent to put Dr. in an alias field.

When we use aliases, we’re trying to get at every possible thing that a user could search for. If they happen to search for "Dr." in their entity search, they could find that Ivan Campiato. If they just put the last name, the same person would show up, or if they had used a di􀆯erent name for a period of time another Ivan Campiato would be returned in the search results. So, it’s not quite as structured as NACO, but it’s meant to be broad.

Is OCLC Meridian already integrated to OCLC WorldShare or is it a stand-alone product of the OCLC?

It is both. It is a standalone product. There’s a scenario where you maybe don’t have any WorldShare products but need to aggregate things together by an entity. However, maybe you have OCLC for cataloging and you wanted to use Meridian to manage entities. That can be done too. Record Manager and Connexion both have a lookup. We’re creating some integrations so that if you’re cataloging using Record Manager and you want to create an entity right in that flow, we have that integration there. We also have APIs, so it’s possible to consider integrations with other systems as well, but it can be certainly used by anyone using any systems.

In Record Manager, there’s a way to create a WorldCat Entity from your working in a MARC bibliographic record. There’s a function that if you’re a Meridian subscriber, you can create the WorldCat Entity from that, and also you can look up a WorldCat Entity URI. Connexion doesn’t have that kind of functionality to be able to create the description directly from Connexion to go into Meridian, but you can look things up.

I use Connexion to create NACO authorities but I don’t see an option to do that in Meridian. When I sign into OCLC services and click on Meridian, it takes me to WorldCat Entities and the only option is to search.

Meridian is a subscription service. You can go into WorldCat Entities to view the data that has been created and managed, and then subscribers can create and edit things.

Apologies if I missed, but is there any part of the EMCO record that contains the title of the work by Campeotto being cataloged?

EMCO is a broader program and within it we have ISNI, Wikidata, and WorldCat entities. So the view that we're taking in this particular session is the Worldcat entities, which is part of the program. When you’re looking at a WorldCat Entity, there are some related works and that’s where you could put works associated with the person. For any given property, you can put a reference to where you found that information. In WorldCat entities, there isn’t a property that indicates that this originated from this particular work, but you could certainly use the bib record as a reference for a given claim property.

How are the FAST subject terms are prepopulated?

In the Meridian interface, if you were creating an entity, there’s an edit field. You can indicate what property you want to use, like field of activity, and then you can start typing and the dropdown terms are from FAST. You can also add strings if your particular topic or occupation is not found.

Is EMCO pre-populated with data from NAF, or did it start from zero?

We used VIAF as the initial data source for WorldCat Entities. We did a transformation of VIAF data and brought it into WorldCat Entities. Now recently, we have begun picking up through the Library of Congress linked data. As a NACO record is created, we will bring in the basic data from that record if it meets certain criteria. We don’t do a full synchronization of all of the data that is in LCNAF because WorldCat Entities exist as a data set on its own and we don’t want any of those changes to overwrite intentional data or changes that have been made by users.

If I understood Anne correctly, what collocates entities using Meridian is a URI as opposed to the form of the heading in LC authorities. Does that mean that the heading may appear in different forms in bib records as long as they are linked by the correct URI?

That’s the vision. Not all bibliographic records have URIs in them right now, but I’m sure you have seen them in the subfield $1s. If you go to WorldCat.org and find a little person icon next to one of the contributor names, you can click on that and it says, “find works by this person.” That does a search on the identifier. The intention future facing is that we can have some precise searches across names in di􀆯erent languages with certain varying forms when we have the URI, but the coverage, we’re still working to build full coverage of URIs associated with names across WorldCat.

Do you have to have a Meridian subscription to be in the early adopters?

You don’t. You can join and be part of the group that is building some of the recommendations.

In NACO one does not normally create a record for a heading unless it is used in a WorldCat record. Does Meridian require that the entity be used in a WorldCat record?

WorldCat Entities is meant to support the breadth of library activity. One component of that is cataloging, and then we have things like local collections of talks or thesis and dissertations or digital digitized special collections materials that may not be cataloged. It’s helpful to have these entities because you can create connections between them. It’s helpful to be able to have an identifier to collect all of the work. This is meant to support broader use cases.

Is Meridian linked data available in ContentDM?

Users can add WorldCat Entities URIs to metadata in CONTENTdm if they choose. OCLC’s goal to bring linked data functionality into both existing and new tools, and we are evaluating di􀆯erent products and integrations that would be helpful to users.

How are entities in Meridian related to LOC entities/linked data?

WorldCat Entities, created and managed in Meridian, represent people, organizations, places, and events with descriptive data that disambiguates and connects entities. The Library of Congress similarly provides authority file data as linked data maintaining consistent name strings with select disambiguating information. These systems are complementary and WorldCat Entities explicitly links to LC linked data through a sameAs relationship, which can be found in the linked data identifier section of each entity record. This connection allows both systems to work together within the broader linked data ecosystem.

Since NACO has been moving toward description rather than just disambiguation, and NAF records come with URIs, has there been any conversation about removing barriers to NACO participation (such as adding a more user-friendly interface) rather than building a new system?

The term “NACO Lite” has been used to discuss the concept of lowering barriers to NACO participation in a variety of ways, including the expertise needed to create MARC records that conform to RDA and PCC standards. Those barriers to membership are not related to any specific system used by catalogers. It’s important to note that the EMCO program is not a replacement for NACO membership, and EMCO participants can use Meridian, ISNI or Wikidata to participate in EMCO.

Thinking on the bib creation end of things. At some point (like when the system is more fully developed and adopted), would using a URI make it unnecessary to use a specific form of the heading?

Yes, one of the benefits of linked data is that multiple human-readable text stings (including traditional catalog headings) can be part of the description of an entity. Libraries would use the URIs for the linked data descriptions, but end user could see human-readable text stings associated with those URIs. In other words, English-language catalogs might display “United States” and Spanish-language catalog might display “Estados Unidos” but the same URI refers to this place.

19 March 2026

Does “lives in” have a URI?

Yes. So every relationship in a linked data world also has a URI, so it’s a land of URI’s. It would be: Anne has a URI, lived in has a URI, and Atlanta has a URI.

It’s a kind of authorities, is it so?

There is a bit of a di􀆯erence between the authority data, kind of that authoritative string that is used and kind of the primary, one of the primary purposes for the name authorities, but the entity data is meant to have, in a given registry, a single representation of a given person or place. Now, that same person could appear in another registry, and it can be helpful or useful to have di􀆯erent registries or contexts where this data is managed because it’s possible that there’s a di􀆯erent focus from one registry to the next. What’s important is that we find ways to connect these di􀆯erent entities so that we can make the best use of all of the data that exists across these di􀆯erent registries.

Can we only use already created Entities during cataloging? Not to create new entities?

Right now, WorldCat entities are available to use within Connexion and Record Manager. We have that integration to insert and add a WorldCat entity, or when you control some headings bring in that subfield $1. Now folks that have OCLC Meridian can create and edit entities.

Is VIAF also a registry?

It is not for this early adopters program, but those that have interest in VIAF could potentially get together. Now, VIAF in its purpose is a little bit di􀆯erent in that it’s not really an active place that people can edit data because it is authority based data. I don’t know that it would totally fit into entity management.

Why does the system mandate fast vocabulary when so many librarians are more familiar with LC?

Meridian was re-released maybe two years ago, and we started with FAST as the vocabulary for things like field of activity or occupation and knew that over time we would learn what vocabularies would be helpful and important to people. So the idea is to add those in where we learn from folks what they’d like to use. In addition, folks can enter string values where there isn’t a value in any authority, for maybe a specialized occupation or to use something in another language. One of the reasons that that was picked as the first go at things is because in the bibliographic data we also have FAST headings, and so we can make connections between people and places if we’re all kind of using FAST.

Can the “entity” data be edited by everyone in OCLC? Or it function like the current level in OCLC, like Full, National enhanced etc

The entity data in WorldCat entities can be edited by those that subscribe to OCLC Meridian, and beyond that there is not an additional access layer. If you are a user, then you can edit entities and mark them for duplicates. All of the actions taken on an entity, like a create or an edit, have a history so you can see what institution has made a change.

With name changes and resolving duplicates, how does the already established URI play a role in these decisions?

When we’re thinking about duplicates and resolving duplicates, we have a process where once the duplicate is identified and resolved, there’s basically a cleanup process that the keeper entity absorbs that other entity. The keeper identifier is then populated throughout. Any connections that that other entity had, that URI had, in the entity data are fixed as well.

What happens if someone creates an entity in Meridian and then another person creates the same entity in the NAF? Are they automatically linked somehow? Will headings controlled to either the WorldCat Entities entity or the LCNAF record successfully link the same person, even if the strings are di􀆯erent?

A person would have to add the LC identifier to the WorldCat entity in order for it to become part of the controlling process. Right now we don’t have a way to automatically bring that identifier in. Once that LC identifier is added to the WorldCat entity, we know those are representing the same thing in each of those systems.

Is controlled vocabulary required for all metadata related to that person (e.g., a􀆯iliation)?

In WorldCat entities, let’s talk about a􀆯iliation: in the editing app Meridian, the options available to you are other entities. For those types of relationships, it’s asking for other WorldCat entities.

Am I remembering correctly that the ability to edit or add entities in Meridian requires a separate subscription from the standard full cataloging subscription?

Correct.

Can you talk a little more about FAST headings and the creation of entities under EMCO?

Every registry has an editor component. In Meridian, in the editor when you go to occupation or field of activity, as you’re adding that data, it’s doing a search against FAST, and so it’s using the FAST terminology, in that attempt to standardize it within the entity data and then make it useful and connected to the bibliographic data. That’s unique to Meridian. We’re learning about the other vocabularies that would be helpful for folks too, and so those are things that we can consider adding on.

What ensures the longevity of URIs so that they don’t become dead links?

All of the entity IDs, or URIs, live on forever. Even if we resolve a duplicate, that URI for the one that gets merged into another will have a message. If you use it in the UI, it’ll redirect you, and on the computer side of things, it sends a particular message that says this is a permanent redirect to a new entity. We don’t cut it o􀆯 and delete it.

What about Lexicon? As I know it is for creating Authorities? Local authorities or?

Lexicon is more for creating authorities or taxonomy data. To my understanding, it doesn’t have the linked data component to it. It is still helpful for standardizing data across di􀆯erent holdings and subjects, but it is more like an authority. I believe it works sort of like a local authority file, so like a smaller version of the LC NACO authority file that would just be for an institution, and it’s an authority file.

Will it always stay under Meridian or will it possibly be opened up to other application options like Connexion (or is Meridian eventually replacing Connexion?)?

Meridian is not the tool that would replace any of our bibliographic editors. It will stay focused on contextual entities. For our cataloging tools, we want to provide ways to add existing URIs to the data or create data where it makes sense for you. You can also create them from inside of Record Manager. It will not ever be a replacement for a cataloging tool.

How are different bibliographic identities linked to each other, e.g., Jean Plaidy/Victoria Holt/Philippa Carr?

We have relationships. Those are all properties that are outlined in the WorldCat ontology. We have a pseudonym property that can relate from one person to another, but one thing is it can be a choice as to people know this is the pseudonym and it’s part of the author’s known identity versus areas where you may not want to connect a person to a pseudonym for privacy reasons.

To add to this, doesn’t using only FAST possibly limit the accessibility of terminology used? If it is having some of the same issues that LCSH does, it may not include terminology that properly describes identities of persons.

Meridian is a new product and, based on users’ feedback, we may add other vocabularies in the future to describe WorldCat Entities. PCC has instructions and guidance for describing people, as well as privacy considerations. This is one of the issues that is also relevant for EMCO discussions of best practices.

13 January 2026

Now that the Library of Congress has decided to stop using form subdivisions, will the authority records for these subdivisions continue to be available in Connexion?

Yes, the Library of Congress has said that the authority records for form subdivisions will continue to be a part of LCSH so they will continue to be available within Connexion and Record Manager since we have a copy of the LCSH authority file. For more information, see the FAQ on omission of $v in LCSH and expanded use of LCGFT, which is provided on LC’s website.

Why aren't those "see also" references in the authority record? (referring to " $x Study and teaching", See $x Instruction and study for music...)

That is an excellent question, but I don’t know the answer. I am sure LC considered doing it as some point, but you would have to ask them why they did not.

So, no one has done the programming to change all of the technically incorrect but still authorized headings? How many years has this been going on? Why bother with it being incorrect if you're going to allow the incorrect ones to authorize?

It’s important to remember that controlling is simply matching authority records. It is in no way a validation of the content or construction. We covered this in our September 2025 Askqc office hours, WorldCat records revealed: The secret lives of validation and controlling. These SHM instruction sheets are very specific and contain many exceptions. It would be incredibly challenging and maybe impossible for a program to be able to recognize when, for example, a heading represents a musical instrument and therefore the free-floating subdivision $x Instruction and study is correct.

Note that “authorized” is a different concept than “controlling headings.” When the SHM talks about a subdivision being “authorized,” that means it is being used appropriately according to instructions. Controlling headings links access points in bibliographic records to authority records.

Are there any plans at OCLC to update this programming in the next few years?

The complex nature of SHM instructions makes it unlikely that we would be able to incorporate them into our controlled headings logic.

Looking ahead to linked data, what parties are working on designing how this type of work will be supported in the future?

There are lots of efforts to support subject access with linked data. Our FAST (Faceted Application of Subject Terminology) project supports a more linked-data friendly approach to subjects because it breaks down what would be a multi-faceted heading in LCSH into separate terms, aka facets. FAST is derived from LCSH, but it is a separate vocabulary. It is easier to apply, but you lose some context. In the future, we may be able to get some of the contextual linking back.

I don’t have an example handy, but I have seen FAST headings with subdivisions.

Yes, there are some FAST headings with subdivisions. One example is geographic headings, e.g., “651 7 Ohio $z Columbus. $2 fast $0 (OCoLC)fst01205076.” There are also some topical terms with subfield $x subdivisions like “650 7 $a Football $x Defense. $2 fast $0 (OCoLC)fst00931378.” You won’t see two different facets like topical and chronological together though.

Does "$x Instruction and study" imply the sort of education that involves actively instructing someone how to do something (playing an instrument) versus the sort of education that involves teaching about a topic? (the history of an instrument) Is that why there's a different subdivision used?

I don’t know the history of this subdivision, but I doubt there was that kind of thought put into it. I suspect what happened is that there were existing subdivisions used under musical instruments, then later there was the free-floating subdivision $x Study and teaching created. At the time this happened, LC might have thought it was too difficult to change all the music subject headings to $x Instruction and study. They might have still been using cards at that point. Also, music cataloging is a specialized area at LC so all the resources about the study and teaching of music would go to the music cataloging team, and they knew which subdivisions to use because they cataloged music all day.

Does anyone know if LC is planning on expanding the LCGFT in light of their decision to stop using form subdivisions? Ex- Children's fiction

LC has said they are reviewing existing form subdivisions and LCGFT to identify gaps. This is discussed in their FAQ document.

22 January 2026

Can you clarify the use of subfield $y (e.g. “XX century”)? Can it only be used after $x History?

Sometimes $y chronological subdivisions are used with $x History (as a free-floating combination), but some areas—such as Art, Philosophy, and Architecture—use the free-floating chronological subdivision without $x History. In other cases, a particular $y is only allowed when established as part of a specific string (i.e., not free-floating). Consult SHM H 1647 (History) and the relevant pattern-heading sheets to determine when to use $x History + $y vs. $y alone.

What is the rationale for treating “classes of persons” separately from ethnic groups, races, and nationalities?

This question is referring to the fact that SHM has these two instruction sheets for free-floating subdivisions: H 1100 Classes of Persons and H 1103 Ethnic Groups (also covers races and nationalities). The SHM does not explain why this was done, so I do not know for certain. My best guess is that there are some instructions at the beginning of each sheet that are specific to one and not the other, e.g., H 1103 has instructions on nationalities that don't apply to classes of persons, and there are some subdivisions in H1103 that are not in H 1110 like $x Colonization. If $x Colonization were in H 1100, the instruction sheet would need to explain all the headings it cannot be used with. But, only someone at the Library of Congress could provide a definitive answer to this question, and they might not know since whomever made this decision no longer works there.

Is there an effort to put “See also” references (e.g., from Study and teaching to Instruction and study in music) directly into subdivision authority records?

Not that the presenter is aware of. Such cross-references would be helpful, but LC has not implemented them in subdivision authority records.

What guidelines exist for entering years in $y—when to use decades, centuries, or year ranges?

Specific chronological subdivisions must be either (a) allowed as free-floating (e.g., via H 1095 for general use or via pattern-heading sheets), or (b) established as part of an authority string. For general history, free-floating options include History—To 1500, History—16th century, … History—21st century. Additional or different chronological breakdowns may be specified in relevant pattern-heading instruction sheets (e.g., Art). If not free-floating or covered by a pattern heading, the chronological subdivision must be established.

Should the 680 scope notes be used when deciding how to apply a subdivision?

Yes—680s are useful, but they often don’t include all constraints. For example, the 680 for Study and teaching looks permissive, but the applicable instruction sheets (H 1095 and then H 2110, which points to H 1160/H 1161 for music) change the choice to Instruction and study for musical topics. Always follow the SHM instruction sheets referenced (e.g., in authority record 073 fields) for the final decision.

Where can I find the relevant documentation in ClassWeb Plus?

In ClassWeb Plus, go to the Cataloging Policy Documents section, open the menu, and then the Subject Headings Manual (SHM).

Is there a link to “pattern heading” information on the QC website?

Yes: Pattern headings for LC subjects on the OCLC help site.

Can you show an example of an instruction sheet?

The panel shared SHM H 1095 (General free-floating topical and form subdivisions) as a PDF example.

Could you talk briefly about the LC deprecation of $v form subheadings, especially as it would affect OCLC records?

Yes, but it's important to be clear about what LC is doing. They are not removing form subdivisions from their authority file. Because we have a copy of LC's subject authority file, including those subdivision authority records, we can continue to control them as part of controlled headings for the many existing and new WorldCat records with LCSH form subdivisions. LC has said they will stop using form subdivisions in bibliographic records for their current cataloging done with the Marva BIBFRAME editor. LC records that are not created with the Marva BIBFRAME editor, like their serial records, will not be impacted by their decision. LC has also not said anything about removing form subdivisions in their existing bibliographic records in WorldCat.

Sometimes, existing WorldCat records are replaced by new LC records. (This depends on the encoding levels of the records and other factors.) Therefore, it is possible that some WorldCat records with LCSH form subdivisions may be replaced by new LC records without them. We are monitoring the situation and will assess the impact on WorldCat data and discovery when LC implements its new policy. If future adjustments to OCLC systems or processes become necessary, we will communicate those changes to our members.

Does OCLC plan bulk corrections/updates for misapplied subheadings?

We do accept suggestions for data quality projects if you would like to email us with a specific suggestion. We are currently working on updating bibliographic records to change $x Study and teaching to $x Instruction and study after some musical instrument headings. However, because there are so many musical instrument headings in LCSH it is hard for us to find all the potentially incorrect the bibliographic records.