This article presents a research on the chronological distribution of coins, excavated in Bulgaria and published in the Proceedings of the (Bulgarian) Archaeological Society (PAS) during the period 1910-1920 and in the Proceedings of the (Bulgarian) Archaeological Institute (PAI) during the period 1921-1959. We denote this set of coins by CFBP-1910-1959 (Coins Found in Bulgaria and Published 1910-1959) and construct their Chronological Distribution (CDCFBP-1910-1959) and its graph using computer technology (the package MS Excel in our case).

The old coins excavated in a given country are important source of information about its past. They bring information about the welfare of its population, its trade centers, trade relations with other countries, religion, the names and the titles of its rulers, etc. The well known specialist in Byzantine history A. P. Kazhdan wrote:

“The coin, that has been a means for trade in ancient times, in the hands of historian nowadays becomes evidence of the intensity of the trade: the abundance of coins excavated in a given old city can serve as a serious argument, which proves the existence of manufactory in this city. The chronological definiteness inherent to the numismatic material advantageously distinguishes it from many other kinds of historical sources. And at last the coin finds give a mass material, which is more or less uniformly distributed over the time and the space.” (Kazhdan, 1954, p. 166)

We ought to add to the above that the coins are a reliable “dating element” for other archaeological evidence, found together with them. Therefore the old coins are object of special interest not only for numismatic collectors, but also for archaeologists and historians. Is it possible the quantity of the found coins, struck by a given ruler, to give us certain objective information about the monetary circulation during his reign, and thus about the intensity of the trade relations in his country and between it and other countries?

In the cited article (Kazhdan, 1954, p. 166) A. P. Kazhdan gives positive answer to this question with some reservations of little importance. To obtain, however, the needed information one has to handle in a proper way the data about the coin finds. We do this here, applying the modern computing technology to the ideas of A. T. Fomenko about his concept of the so called “volume function” introduced in Fomenko, 1981a and Fomenko, 1981b, pp. 211-212 (for more details see Fomenko and Rachev, 1990, pp. 187-206).

Our sources of information are the Proceedings of the (Bulgarian) Archaeological Society (PAS) during the period 1910-1920 and the Proceedings of the (Bulgarian) Archaeological Institute (PAI) during the period 1921-1959. In a certain sense PAS has been continued by PAI. For the studied period these journals were the only scientific ones, in which regular reports of Bulgarian archaeologists on excavations and finds in Bulgaria have been published. We have extracted data about more than 280.000 old coins grouped in more than 1.050 finds from these journals. The high scientific verification of our data is based on this approach. There are no repeatedly reported finds, which might occur if for example additional information from newspapers or other journals is used.

The rulers, who struck the respective coins are usually given in the reports of the archaeologists. We use their time of reign for dating the coins. For about 20% of the coin finds there are given neither the rulers, nor some approximate dating – for example “there were found 70 gold Byzantine coins. I saw 4 of them”. Such a piece of data is ill suited for use in this study because of uncertainty of the dating and therefore have been left out.

The data about the quantities of the coins for several coin finds in PAS and PAI is given in kilos or pots. We assumed that each kilo equals 50 coins and that a pot contains approximately 500 coins and below we will stick to this dimension. Admittedly there were also found other ancient coins during this 50-year period (1910-1959), which have been published neither in PAS, nor in PAI. Many times some coins were found by treasure-hunters and then sold to private collectors without the civil authorities or any scientist to be informed about it. But the “split” of coins to “identified and published” and to “not published” has a random character. Furthermore the quantity of the identified and published coins is very large (more than 280.000 items) and for this reason it will be plausible to suppose that our sample is representative for all the coins found in Bulgaria. Therefore we assume that our conclusions for CFBP-1910-1959 hold with a high probability for all coins found in Bulgaria.

Basic time unit: twenty years. We fix intervals of time in the form (1200, 1220], (1220, 1240] as time units, which bear chronological information. Our choice is conditioned mainly by the following reasons:

1. The length of such interval and its ends are integer numbers ending in zero, which is convenient for our purposes.
2. The length of such interval is close to the average length of the reign of a ruler.
3. The necessary round off has no significant influence on the accuracy.

Note. In Tabov et al., 2003 we have used unit intervals of length 10 years. Our examination showed that the transition to units of length 20 years slightly influences the accuracy but the speed of entering the data into the computer becomes almost twice quicker.

Let us now illustrate the details of the construction of the Chronological Distribution of CFBP-1910-1959 by some examples. Coins’ dating: as usually, we attribute the coins to the ruler who struck them. If a coin has been struck by Tsar Ivan Shishman (1371-1393), then we attribute it to the interval (1371-1393).

Round off and transition to basic time units: in the case above we “put” the coin in the interval (1380-1400). We associate a special function with this coin, the so called individual unit-step function (IUSF), which equals 1 in “its” interval and 0 out of it. The graph of this function is shown in Fig. 1:

The so constructed CDC is a function of time only, because the set of coins CFBP-1910-1959 is fixed and the summing up includes all the coin’s IUSF. The number of the IUSFs is such that the use of computer technology becomes expedient. The MS Excel package can be used conveniently for adding them up and drawing the graph of their sum CDC. More details are given in Tabov and Tabova, 2004, Tabov et al., 2005, pp. 103-107, Tabov et al., 2004a, pp. 99-104 and Tabov et al., 2004b, pp. 257-261.The graph of the derived CDCFBP-1910-1959 is very convenient to make analyses and to draw conclusions.

Since the coins’ excavation is a random process, it is natural to expect that after entering large enough quantity of data the shape of the graph of the CDC would become stable. This expectation proved to be true. To show this fact we consider the graphs in Fig. 3 and Fig. 4, which present respectively the Chronological Distribution of parts of CFBP-1910-1959, namely the Chronological Distribution of the coins found during the first 25 years of the studied period (published in Tabov et al., 2005, pp. 103-107), and the Chronological Distribution of the coins found in 1910-1950 (published in Vassilev et al., 2005).

It might be checked that the sequence of the CDC for the first 26, 27, …, 40, …, 49 years “approaches” the CDCFBP-1910-1959.

On the basis of this observation it is natural to expect that:

We will consider four anomalies, which leap to the eye and are the most significant in our view.

1. About 60% of all found coins belong to the interval (-400, 400].
2. The amount of coins dated to the period (400, 1050] is negligible.
3. The percent of the found Turkish coins from the period of the Turkish slavery, i.e. minted after 1400, is very small, which looks very strange for the epoch of 16th -18th century.
4. The periods (200, 260] and (1180, 1240] are represented with great amounts of coins.

The first two anomalies show that the mistakes in the dating of some old coins are quite probable. We will consider at length the second one.

In his research on Byzantine towns, published in Kazhdan, 1954, pp.164-188, A. P. Kazhdan studied every peace of data on the old coins as objective archaeological material. He used sources of information of two types:

1. Catalogues of big museums.
2. Reports on coin finds (excavations and collective finds).

Kazhdan wrote: “The catalogues of the big collections contain descriptions of hundreds of coins and provide opportunities to present more or less tentatively the minting intensity during one or another period of the Byzantine history” (Kazhdan, 1954, p. 166).

In other words, without restriction in the limits of Byzantium and its history, the analyses of information of big enough and randomly chosen quantity of coins delivers opportunities to model the intensity of the monetary minting in the past. On the basis of the data on the Byzantine coins in the collection of the British Museum, published in (Wroth, 1908), Kazhdan created the table given in Fig. 5:

Considering further the generalized data of finds of Byzantine coins Kazhdan makes the following conclusion: “Considering the data in these catalogues one may suppose that at the border between 7th and 8th century Byzantium had entered in a period of significant economical alterations, which found expression in diminishing the intensity of minting.

Its graph is shown in Fig. 9:

Thus Kazhdan convincingly shows the existence of a “monetary crisis” in the Byzantine history. It comprises the period 8th-9th century.

Kazhdan suggests an answer, which likely follows from his research: during 8th-9th century there were decays and degradation of the towns. According to him all Byzantine towns had stopped to exist or became villages and some towns like Constantinople, Thessalonica and Nicaea survived but had lost their leading economical status.

Kazhdan wrote in his article (Kazhdan, 1954, pp. 164-188):

“The material studied by us do not allow us to agree with the statement, wide-spread in the bourgeois historical literature, that Byzantine Empire has always been an urban country. Indeed with the fall of the slave-holder way of manufacture in the East Roman Empire gradually disappeared almost all slave-holder Antique polises on Balkan Peninsula and in Asia Minor: the slave-holder polis had collapsed together with the slave-holder way of producing…Only a few towns besides Constantinople survived the fall of the slave-holder society, but they also had lost their economical functions.” (Kazhdan, 1954, p. 187)

The wording here outwardly reminds the attacks on the “bourgeois historical literature” but the core idea is that the towns disappeared. It is in accordance with the concept of the “dark ages” – namely in the 8th, 9th and 10th centuries. It is worth to pay a special attention to the “exceptions from the rule”, i.e. to those towns, for which Kazhdan wrote that they had preserved their urban appearance. It makes at once the impression that for none of them Kazhdan gave any data about the number of the coins found in the respective town or in the surrounding area.

Considering the catalogue data of the British Museum and the collection of count Tolstoy we see that the number of coins in the “survived” towns, which came to us from 8th and 9th centuries, is too little. Hence the quantity of the found coins with a high probability is small not only in the survived towns. Therefore no objective data is visible except probably of some written sources, which might allow making such “exceptions”.

All this creates the feeling that Kazhdan had tried to soften the clear results from his research and thus partially to hide the contradiction between them and the generally accepted historical picture of the epoch 8th – 10th century.

Several years after the publication of the article of Kazhdan another soviet historian – I. V. Sokolova – suggested other explanation for the lack of coins during the period mentioned in her article (Sokolova, 1959, pp. 50-63) and tried to solve the problem stated by Kazhdan. First of all we ought to point out that she had understood very well the meaning of the coin finds for the work of the historians and especially for the decay of the towns during 8th-9th century. Her position on it is expressed in the introduction of her article:

“Coins are very rich, many-sided and interesting historical source. Designed for serving as a tool for trading they are in position to reflect to a considerable degree the state and the level of the economical development of a country during given period. Therefore the investigations of the coins themselves, of the monetary system, of the distribution and the content of the collective coin finds, random coin finds and the coin finds in cases of archaeological excavations as well create opportunities to make clear the picture of the monetary circulation, characterizing the economical centers and regions, the inner trade paths, international trade relations, etc.” (Sokolova, 1959, p. 50)

But this acknowledgement of the role of the coin finds does not mean at all that she was ready to accept of Kazhdan’s methodology and conclusions. This is clear from her words further down: “One has to take into consideration that the numismatic material can be valuable only when it is studied from all the sides, with correct methods and comparing it with the data from the written sources. The one-sided selective usage of it may lead to wrong conclusions.” (Sokolova, 1959, p. 50)

This gives us the opportunity to see that Sokolova has seen well the contradiction between the conclusions of Kazhdan and the generally accepted picture of the past, and in final account with many pieces of information from the narrative sources. Sokolova puts the specific understanding of the concept collective coin finds in the root of her disagreement. Since the coins are frequently found “many on the same place” in sets of some or several hundreds or even thousands (namely for such coin sets she uses the term “collective coin finds”), she starts her criticism to the work of Kazhdan with the note:

“Therefore Kazhdan is not right considering a collective coin find not as an overall memorial from the epoch of its burial but as a randomly accumulated material. Decomposing the finds in parts of coins from same centuries and then summarizing the respective parts of all the collective coin finds, A. P. Kazhdan compares the results thus obtained considering them as a measure of the intensity of coins’ minting or their distribution in foreign countries during one or another historical period. At that he doesn’t take into account the size of the minting area and the metal of the coins, which form the collective coin find.” (Sokolova, 1959, p. 50)

Kazhdan really has ignored the material of the coins and also a number of insignificant other things, which influence on the studied problem is relatively small in first approximation. For example for the minting area size Sokolova is not right, because Kazhdan considers a group of towns of the Byzantine Empire, which is representative for the latter and as such it gives completely adequate view on the economic life of the whole country.

But her main objection to the conclusions of Kazhdan is based on the following: “Influxes and refluxes in burying collective coin finds in one or another region depend directly on how much the population of these regions was threatened by war.” (Sokolova, 1959, p. 52) With other words she states that the people in the past had buried their money in the ground because of fear of enemy invasions. But then we must ask: what the people had done to hide money from their relatives – wife, children, and grand children? And from friends and enemies, from thieves and burglars? Put their money in Swiss banks? This concept of Sokolova is unacceptable. In the most cases people hide money not because of war danger but for other reasons.

But even if we suppose that Sokolova is right, then taking into account the frequent Bulgarian raids in Byzantium during the period 8th – 9th century, it is natural to expect that because of warlike danger in today’s South Bulgaria lots of coin treasures have been buried there. However, as we have already seen, the archaeological data shows just the opposite.

So, the objections of Sokolova against the conclusions of Kazhdan are unacceptable. First because they are based on a wrong concept of the “peace” on the Balkans during the period 8th – 9th century, and second because they are not consistent with the logic of the people’s behaviour when they try to hide their money from other’s eyes and hands.

Recently in his article (Romanchuk, 2003, pp. 132-137) A. Romanchuk took a medium position between the described opinions of Kazhdan on one hand and Sokolova on the other about how far the quantities of the found old coins reflect the state of the coins’ minting during different periods of time. He took a note of the fact that some coins were probably in circulation during periods of time, considerably surpassing the period of reign of the ruler, who struck the respective coins. He advances arguments noticed by V. Kropotkin (Kropotkin, 1961, p. 17) that in the archaeological finds of Roman coins from 3rd century there are always coins from 1st and 2nd century as well. If this is true, it points in our opinion to the fact that there are systematic errors in coins’ dating.

We will consider another two anomalies, pointed in § 5. Let compare the quantities of the coins dated to the intervals (220, 320] and (1480, 1580]. From the modern historical point of view of the picture of the past there should be expected that the coins from the second interval are much more than these from the first one. We point two arguments pro this thesis:

1. The Roman reign on the Balkans was in crisis during the first half of the first period. The Barbarian raids lead to stagnation in many aspects of life over the territories of today’s Bulgaria, especially in the field of economics. During (1480, 1580] the territories of today’s Bulgaria were parts of the Ottoman Empire and undergo economical growth. The stories of travellers, who had crossed the Balkan Peninsula on their way to Constantinople, mention caravanserais and dozens and sometimes even hundreds of shops in each town (see for example the recent monograph of L. Klusakova (Klusakova, 2002)), i.e. obvious signs for brisk trade.

2. The second interval is later and the percent of the extant coins should be greater (at least not lesser) than the same percent for the first interval. But from the graph in Fig. 2 it follows that the most intensive monetary circulation through; Bulgarian lands was namely in the period (220, 320]. It means ten times more coins in circulation than those in the period (1480, 1580]. Such levels, so high, are unrealistic, having in mind that a part of Balkan population was slaves during the 3rd century and they didn’t partake in the monetary circulation.

The analysis of the graph of the CDCFBP-1910-1959 (chronological distribution of the coins found in Bulgaria and published during the period 1910-1959) from the point of view of today’s historical science on the picture of the past points a number of contradictions. Their explanations have to be searched in different directions and the most natural of them is to propose that much of the discrepancies are due to certain incorrect dating of coins.

The scale of the anomalies gives us merits to suspect that there are mistakes in the dating of some old coins. For the justification of this hypothesis further investigations are necessary.

Fedorov S., Fomenko A. T. (1986) Statistical Estimation of Chronological Nearness of Historical Texts. Journal of Soviet Mathematics, Vol. 32, № 6, pp. 668-675.
Fomenko A. T. (1981a) New experimental-statistic methods of dating events and applications in the Global chronology of the Ancient world. Preprint Gos. Com. Telev. Rad. 3672, № B07201 (9/1981), Moscow. (In Russian)
Fomenko A.T (1981b) Informational functions and related statistical laws. Thesises of papers presented at the 3-d International Conference in Probability Theory and Math. Statistics, Vilnius, Vol. 2, pp. 211-212. (In Russian)
Fomenko A. T (1990) Methods of statistical analysis of narrative texts and applications in chronology. Moscow University, Moscow. (In Russian)
Fomenko A. T. and Rachev S. (1990) Volume Functions of Historical Texts and The Amplitude Correlation Principle. Computers and the Humanities, 11, pp. 187-206.
Hristova S., Dobreva M. (2004) Some observations on the chronological distribution of mediaeval manuscripts and church items from Bulgaria. In: Mathematics and Education in Mathematics. Proceedings of the Thirty Third Spring Conference of the Union of Bulgarian Mathematicians, Borovets, April 1-4, pp. 214-217.
Kalashnikov V. V., Rachev S. T., Fomenko A. T. (1986) New methods for comparison of volume functions of historical texts. Problems of the stability of stochastic models. Proceedings of a seminary. VINISI, Moscow, pp. 33-45. (In Russian)
Kazhdan A. P. (1954) Byzantine towns during 7th-11th century. Soviet Archaeology, XXI, pp. 164-188. (In Russian)
Klusakova L. (2002) The Road to Constantinople. Sixteenth-Century Ottoman Towns through Christian Eyes, ISV Publishers, Prague.
Kropotkin V. V. (1961) Finds of Roman Coins in the USSR. Soviet Archaeological Institute, Issue G4-4. (In Russian)
Kropotkin V. V. (1962) Finds of Roman Coins in the USSR. Soviet Archaeological Institute, Issue E4-4. (In Russian)
Lazer R. (1980) Die Roemischer und Fruehbyzantinischen Fundmuenzen aus dem Gebiet der DDR, Berlin. Proceedings of the (Bulgarian) Archaeological Institute. Vol. 1(1921-1922), Vol. 2(1923-1924), Vol. 3(1925), Vol. 4(1926-1927), Vol. 5(1928-1929), Vol. 6(1930-1931), Vol. 7(1932-1933), Vol. 8(1934), Vol. 9(1935), Vol. 10(1936), Vol. 11(1937), Vol. 12(1938) Vol. 13(1939), Vol. 14(1940-1943), Vol. 15(1944-1946), Vol. 16(1947-1949), Vol. 17(1950), Vol. 18(1951-1952), Vol. 19(1953-1955), Vol. 20(1956), Vol. 21(1957), Vol. 22(1958-1959).
Proceedings of the (Bulgarian) Archaeological Society. Vol. 1(1910), Vol. 2(1911), Vol. 3(1912-1913), Vol. 4(1914), Vol. 5 and 6 (1915-1918), Vol. 7(1919-1920).
Romanchuk. A. I. (2003) On the problem of the usage of numismatic data in studying the Byzantine town during the period of the “dark ages”. In: The world of Alexander Kazhdan, Alateya, Sanct-Petersburg, pp. 132-137. (In Russian)
Sokolova I. V. (1959) Finds of Byzantine Coins as a source of the Byzantine History in 7th -11th century. Viz. Vremennik, XV, pp. 50-63. (In Russian)
Tabov J. (2003) Chronological Distribution of Information in Historical Texts. Computers and the Humanities, 37, Number 2, May, pp. 235-240. 18. Tabov J. and Tabova L. (2004) Searching for Dark Ages. Storiadelmondo, № 28, 19 July. [http://www.storiadelmondo.com/28/tabov.searching.pdf]
Tabov J., Vassilev K. and Velchev A. (2003) A mathematical model of monetary circulation in Medieval Bulgaria. Storiadelmondo. [http://www.storiadelmondo.com/14/tabov.monetary.pdf]
Tabov J., Vassilev K. and Velchev A. (2004a) Mathematical modelling of monetary minting in mediaeval Bulgaria. Review of the National Center for Digitization, Belgrade, 4, pp. 99-104.
Tabov J., Vassilev K. and Velchev A. (2005) Chronological Distribution Of The Coin Finds In Bulgaria Reported in the Scientific Literature For a Quarter Century (1910-1934), Review of the National Center for Digitization, Belgrade, 6, pp. 103-107. [Published at Chrono-Logic
22. Tabov J., Velchev A., Dobreva M., Sotirova K. (2004b) Chronological distribution of the Bulgarian mediaeval manuscripts preserved in Bulgaria. In: Mathematics and Education in Mathematics. Proceedings of the Thirty Third Spring Conference of the Union of Bulgarian Mathematicians, Borovets, April 1-4, pp. 257-261.
Tolstoy I. I. (1912-1914) Byzantine coins. Sanct-Petersburg. (In Russian)
Vassilev K., Velchev A. and Tabov J. (2005) Finds of old coins in Bulgaria for fourty one years (1910-1950)], Sbornik of Papers in New Chronology, Issue 2, 07 February. [http://new.chronologia.org/volume2/41.html] (In Russian)
Wroth W. (1908) Catalogue of the imperial Byzantine coins in the British Muzeum, London.