A historical image is both an image and an object with a history: made by someone, for a reason, with a particular camera, then copied, sold, published, kept or forgotten, and given several captions on the way. Beyond Yesterday follows the whole path, from the Manakia brothers’ photographs through their archives to the landscapes in which they were made. This page says what the project counts as evidence, what it counts as interpretation and when it leaves a question open.

1. The archive as a set of objects

One scene may survive as a glass negative in Bitola, as a print in a family album, as a postcard with a commercial caption, as a newspaper reproduction, as a copy in a museum and as a digital scan. These are not six identical sources: the glass keeps technical traces, the album an order, the newspaper a date, the back of a print may carry a name, and a scan may crop the margins. The project therefore keeps separate records for the scene shown, for each surviving object, for each edition and for the digital reproduction. The first question is always material: is the negative an original or a contact copy, was the print cropped or retouched, can the handwriting or the studio stamp be dated? An archive description is part of the object’s later history and is not the same as the event shown. The institutions are presented in Archival Sources.

2. How an attribution is made

An attribution starts from the object: format, marks, inscriptions, the caption in earlier catalogues, condition and the holder’s identifier. Then independent witnesses are sought (maps, newspapers, school registers, letters, building histories, dated theatre programmes, other photographs) that can confirm or contradict the caption. A face, a mountain skyline, a costume, a shop sign or an architectural detail is compared with several possible identifications, and evidence against is sought as persistently as evidence for.

Observation is kept apart from inference. “The sign reads The Liars” is an observation; “this was the visit on which Camera 300 was bought” is an interpretation and needs more than the sign. The reasoning is published with its sources, and the old catalogue label is kept even where we explain why it looks wrong. New evidence can change an identification, so every major revision stays dated and named.

A famous face calls for extra caution, because fame makes a resemblance look more convincing than it is. A caption can pass from a memory into an archive inventory and from there into dozens of publications without the image ever being compared with independent portraits. When we publish an identification we show the original plate, the old caption, the whole image (not a cropped face), the comparison images with their provenance, the method and the named expert. A forensic comparison is valuable evidence but does not replace historical context. The case of Mustafa Kemal is in the Journal.

3. The 1905 / 1907 question

The first Manakia film is conventionally dated to 1905, a date repeated by the Cinematheque of North Macedonia’s chronology and by many reference works. It is not an invented date but a historiographical tradition resting largely on later recollection.

Beyond Yesterday dates the purchase of Camera 300 and the first film to 1907, in agreement with Emil Țîrcomnicu’s argument from documents. The new evidence is concrete: a broken glass negative kept in Bitola shows Ianaki outside the Criterion Theatre in London, and the theatre’s programme shows that The Liars ran there from 13 April to 15 October 1907. The place is near the firm of Charles Urban, supplier of the Bioscope camera. The first moving images, among them the film of his grandmother Despa at Avdela, come after the purchase.

The wording stays cautious: traditionally dated 1905; after the research of Cătălin D. Constantin and Emil Țîrcomnicu, the camera was bought and the earliest film made in 1907. A dated bill of sale or a contemporary witness would be decisive. In full: “The man in front of the theatre” in the Journal, Țîrcomnicu, Ioan D. Manakia în documente (2024), and Dimitar Gjorgievski’s study of Camera 300 (2022).

4. Several kinds of date

Photographic archives usually keep one date field for different things. Here we distinguish the date of exposure (the most important, but sometimes uncertain), of the event (a dated visit can give a firm interval), of the print (it may be decades later than the negative), of publication (a newspaper of 1903 gives an upper limit for the exposure), of entry into a collection and of modern identification. For hard cases the earliest and latest possible dates are given; “before 20 March 1903”, “1898–1900” or “photographed by October 1907” are more accurate than an unjustified precise year.

5. Film: titles are not productions

The brothers’ films exist in several forms: a camera roll, a modern edition that extracts a few metres and gives them another title, a compiled film with added intertitles, DVD chapters catalogued separately. A figure about films must therefore be clarified first: does it count camera rolls, original titles, surviving elements, archive records or edited chapters? Are lengths given in metres or in running time, and at what frame rate? Is the surviving object a camera negative, a positive print, a duplicate or a restored master? Are the intertitles original or added later? The case of the 1911 Romanian delegation shows why it matters: the film edited in Romania, Excursie în Macedonia turcească, has a connected itinerary, while modern records may split the parts on Bitola, Gopesh and Resen. Different cataloguing practices can create the appearance of several unrelated works; the event itself remains of 1911 (Marian Țuțui, “The Visual Heritage of Manakia Bros.”).

6. Comparative photography

The procedure itself (candidate place, fixed points, point of view, recording of change) is described in Field Method, and the case studies from Laista, Korçë, Plovdiv, Sliven and Bitola are in the Journal. What matters here is its use as verification: a wrong caption gives way before the ground, while local memory must itself be attributed and corroborated. A convincing montage is not a geometrically exact reconstruction, and we say so when the point of view is only approximate.

7. Names, identities, belonging

The Manakia brothers lived in a region where language, school, confession, political loyalty, citizenship and everyday belonging did not necessarily coincide. The same person could grow up in an Aromanian-speaking family, attend a Romanian-language school, work in an Ottoman town and be claimed later by several national histories. Even their names circulate in several forms (Ianaki, Ienache, Janaki, Yannakis, Giannis; Milton, Miltos; Manakia, Manaki, Manakis, Manaqi). The form in a source shows the language and political setting in which it was written and is no reason to exclude the others. Here we generally use Ianaki and Milton Manakia; the bibliography keeps the printed spelling.

The question whether they were “Greek”, “Romanian”, “Macedonian”, “Albanian”, “Turkish” or “Yugoslav” pioneers easily replaces research with a division of symbolic property. It is more exact to describe separately the place of birth, the languages, the trades, the institutional ties, the subjects photographed, the historical citizenship and today’s custody. Identity mattered much in the late Ottoman Balkans, and some photographs were made in institutions with explicit national programmes; the hostility of part of the Greek press towards Ianaki and the context of the Romanian school belong to the same history. But no single label describes the photographs and all the people in them. Readings: Siokis (2024), Guštin (2023), Özen (2010, 2012), Grgić (2020, 2022), in the Bibliography.

8. Milton’s year of birth

The literature does not give the same year: 1882 appears in institutional biographies and in a 2022 anniversary publication of the State Archives in Bitola, 1880 in recent curatorial material, among it the 2023 programme of the Albanian Ministry of Culture for the Marubi exhibition. Beyond Yesterday uses 1880, reflecting the project’s present reading, and flags the difference: Milton Manakia (1880–1964; 1882 is also widely given in the literature). Until a primary civil, church or school register is found and published, it remains an open documentary problem.

9. Degrees of certainty

The research notes use four degrees, so that the reader can tell what is observed from what remains provisional.

• Documented: independently dated photographs, signed or dated documents, catalogue references checked against an original or an authorised transcription.
• Firm identification: several independent kinds of evidence support a person, a place or a date, and the argument and comparisons can be checked.
• Provisional: a plausible attribution that still depends on missing evidence.
• Open question: a contradiction or gap worth keeping in view rather than covering with an attractive story.

The language follows the degree: “probably photographed at…”, “may show…”, “previously catalogued as Plovdiv; proposed identification: Sliven”. “Certain” is not an archival method, and neither is “legendary”.

10. Artificial intelligence and the archival image

Enlarging, adjusting contrast, overlaying two frames and searching for images can bring out details that are easy to miss, and automatic tools can help find comparison terms in large collections. An algorithmic match is, however, a working hypothesis, not an independent historical witness.

The project also shows animations made with the help of AI in its exhibitions. They can make the subjects of old photographs legible again for today’s public, but they are creative interpretations and are never passed off as original Manakia films. The original image stays accessible, with its citation. Generated movement, and invented sound or speech, are declared. Generated details (a façade, a facial feature, a badge, a sign) are not used as evidence. Technical restoration (a scratch cleaned, a frame stabilised) differs from synthetic creation (an invented missing face). Living people and descendants remain a matter of dignity and sensitive representation. A caption can be simple: “Historical photograph by the Manakia brothers, [archive reference]. Animation made with the help of AI for Beyond Yesterday, [year]. Movement and sound are modern reconstructions.”

11. Rights and digital preservation

Digital access does not erase rights and provenance. A scan from the British Library’s EAP1470 keeps the holding institution in Bitola, the archive reference and the EAP identifier, and funding or hosting a scan does not confer ownership of the negative. Citation formulas and the question of permissions are in How to Cite.

Some questions stay unanswered: why a negative was catalogued with the wrong town, why a film circulates under several titles, why descendants, researchers and film archives remember the same event differently. The research does not aim to take the poetry out of the images or to turn them into illustrations of a closed history. The stronger the image, the greater the need to see what it actually proves; sometimes the background says more than the subject, and a catalogue error can open the way to a discovery. Interpretations are updated when firmer documentary evidence appears.