Frankish Period

|[i]| Frankish Period. |

|[2]| The Revolution in Landed Property Relations under the Merovingians and Carolingians.

The mark constitution remained the basis of almost the entire life of the German nation until the end of the Middle Ages. After having existed for fifteen hundred years, it finally perished gradually on a purely economic path. It succumbed to the economic advances to which it no longer corresponded. We shall have to examine its decline and final downfall later; we shall find that remnants of it still exist today.

But if it maintained itself for so long, this occurred at the expense of its political significance. For centuries it had been the form in which the freedom of the Germanic tribes had embodied itself. It now became the basis of a thousand-year bondage of the people. How was this possible?

The earliest community, we saw, embraced the whole people. To it originally belonged all the land taken into possession. Later, the totality of the inhabitants of a gau, more closely related among themselves, became the possessors of the territory they had settled, and the people as such retained only the right of disposal over the remaining unowned tracts. The gau inhabitants in turn assigned their field and forest marks to the individual village communities—likewise formed of closer blood relatives—while the surplus land then remained with the gau. The same occurred with the original villages when new village colonies were sent out, endowed with land from the old mark of the parent village.

The blood bond, upon which here as everywhere the whole constitution of the people rested, fell more and more into oblivion with the increase of the population and the further development of the people. ||[3]| This was the case first with respect to the totality of the people. Common descent was less and less felt as real blood relationship; the memory of it grew ever weaker; only the common history and dialect remained. On the other hand, the consciousness of the blood bond of the gau inhabitants naturally persisted longer. The people was thereby reduced to a more or less firm confederation of gaus. In this condition the Germans seem to have been at the time of the migrations. Ammianus Marcellinus expressly relates this of the Alamanni; in the tribal laws it still peeps through everywhere; among the Saxons this stage of development still existed at the time of Charlemagne, among the Frisians until the downfall of Frisian liberty.

But the migration onto Roman soil also broke, and had to break, the blood bond of the gau. Even though the settlement was intended to follow tribes and lineages, it could not be carried through. The long marches had not only tribes and lineages, they had even whole peoples thrown together in confusion. Only with difficulty could the blood bond of the individual village communities still be held together, and these thereby became the real political units from which the people was composed. The new gaus on Roman territory were from the outset more or less arbitrary—or determined by pre-existing conditions—judicial districts, or very soon became such.

Thus the people was dissolved into an association of small village communities, among which there existed no, or scarcely any, economic connection, since each mark was self-sufficient, produced its own needs, and besides, the products of the individual neighbouring marks were almost exactly the same. Exchange between them was thus fairly impossible. And such a composition of the people out of nothing but small communities, which indeed have equal but, precisely for that reason, no common economic interests, makes a state power that has not emerged from them, that stands opposed to them as something alien, and exploits them more and more, the condition for the continued existence of the nation.

The form of this state power is in turn conditioned by the form in which the communities find themselves at the time. Where, as among the Aryan peoples of Asia and among the Russians, it arises at a time when the commune still cultivates the land on common account, or at least assigns it to the individual families only for a time, where therefore no private property in land has yet formed, the state power appears as despotism. In the Roman lands conquered by the Germans, on the other hand, we find, as we saw, the individual shares in arable and meadow already as allod, as the free property of the possessors, subject only to the common obligations of the mark. We now have to examine how on the basis of this allod a social and political constitution arose which—with the usual irony of history—finally dissolves the state and, in its classical form, destroys all allod.

With the allod was given not only the possibility but the necessity that the original equality of landed property be converted into its opposite. From the moment of its establishment on formerly Roman soil, the German allod became what Roman landed property alongside it had long since been—a commodity. And it is an inexorable law of all societies resting on commodity production and commodity exchange that in them the distribution of property becomes ever more unequal, the antithesis of wealth and poverty ever greater, property is concentrated ever more in a few hands; a law which, though it attains its full development in modern capitalist production, by no means first becomes operative in it. From the moment, therefore, when allod, freely alienable landed property, landed property as a commodity, came into being, from that moment the rise of large landed property was only a question of time.

In the epoch with which we are dealing, however, agriculture and cattle breeding were the decisive branches of production. Landed property and its products constituted by far the greatest part of the wealth of that time. Whatever movable wealth existed otherwise followed, by the nature of things, the landed property, was found more and more in the same hands as the latter. Industry and commerce had already declined under Roman decay; the German invasion almost completely destroyed them. What remained of them was mostly carried on by unfree persons and foreigners and remained a despised occupation. The ruling class, which in the face of emerging inequality of property gradually formed itself here, could only be a class of great landowners, its form of political rule that of an aristocracy. When, therefore, we shall see how in the rise and development of this class political levers, violence and fraud are active in many respects, indeed seemingly predominantly, we must not forget that these political levers serve only to promote and accelerate a necessary economic process. We shall of course just as often see how these political levers impede economic development; this happens often enough and each time when the various parties concerned set them to work in opposite or intersecting directions. |

|[6]| How then did this class of great landowners arise?

First, we know that also after the Frankish conquest in Gaul a multitude of Roman large landowners maintained themselves, who had their estates cultivated mostly by free or tributary tenants against a rent (canon).

Then, however, we have seen how, as a result of the wars of conquest, kingship among all the Germans who had migrated became a permanent institution and real power, how it transformed the old people’s land into royal domains and likewise incorporated the Roman state lands into its possession. During the many civil wars that sprang from the divisions of the realm, this crown property was continually increased through massive confiscations of the estates of so-called rebels. But as rapidly as it grew, so rapidly was it squandered in donations to the church as well as to private individuals, Franks and Romans, followers (antrustions) or other favourites of the king. When, during and through the civil wars, the beginnings of a ruling class of grandees and mighty men—landowners, officials and military leaders—had already formed, their support was also bought by the partial princes through gifts of land. That all this in by far the most cases were real gifts, transfers to free, hereditary and alienable property, until a change occurred in this respect with Charles Martel, has been irrefutably proved by Roth.*)

When Charles took the helm of the state, the power of the kings was completely broken, but that of the mayors of the palace by no means put in its place as yet. ||[7]| The class of magnates created under the Merovingians at the expense of the crown favoured in every way the ruin of royal power, but by no means in order to subject themselves to the mayors of the palace, their peers. On the contrary. The whole of Gaul was in the hands, as Einhard says, of these “tyrants who everywhere claimed the lordship for themselves” (tyrannos per totam Galliam dominatum sibi vindicantes). Alongside the secular magnates, this also happened on the part of the bishops, who in many regions had appropriated the rule over surrounding counties and duchies, and were protected by immunity as well as by the firm organisation of the church. The internal disintegration of the realm was followed by invasions of the external enemy; the Saxons penetrated into Rhenish Franconia, the Avars into Bavaria, the Arabs across the Pyrenees into Aquitaine. In such a situation the mere suppression of the internal and expulsion of the external enemies could not help in the long run; a way had to be found to bind the humbled magnates, or the successors placed in their positions by Charles, more firmly to the crown. And since their previous power had rested on large landed property, the first condition for this was a total revolution in landed property relations. This revolution is the main work of the Carolingian dynasty. It is distinguished again by the fact that the means chosen—to unite the realm, to bind the magnates permanently to the crown and thereby to strengthen it—ultimately brings about the completest powerlessness of the crown, the independence of the magnates, and the dissolution of the realm.

To understand how Charles came to choose this means, we must first examine the property relations of the church at that time, which in any case, as an essential element of the agrarian conditions of the time, cannot be passed over here.

Already in Roman times the church in Gaul possessed not insignificant landed property, the revenues of which were further increased by great privileges with regard to taxes and other services. After the conversion of the ||[8]| Franks to Christianity, however, the golden age for the Gallic church dawned. The kings vied with each other to make the most donations of land, money, jewels, church utensils, etc., to the church. Chilperic used to say (Gregory of Tours), “See how poor our fiscus has become, see how all our wealth has been transferred to the church.” Under Guntram, the darling and slave of the priests, the donations knew no bounds any longer. Thus the confiscated landed property of free Franks who were accused of rebellion flowed for the most part into the possession of the church.

*) P[aul] Roth, Geschichte des Beneficialwesens, Erlangen 1850. One of the best books of the pre-Maurer period, from which I borrow a great deal in this chapter.

As with the kings, so with the people. Small and great alike could not give enough to the Church. “A miraculous cure of a real or supposed affliction, the fulfilment of an ardent wish, e.g. the birth of a son, deliverance from a danger, all brought a donation to the Church whose saint had shown himself helpful. It was deemed all the more necessary to keep one’s hand ever open, inasmuch as high and low alike held the belief that gifts to the Church procured forgiveness of sins.” (Roth, p. 250.) To this was added the immunity which protected the Church’s property from violence in an age of incessant civil wars, plunderings, and confiscations. Many a little man found it advisable to cede his property to the Church, if the usufruct remained his against a moderate tribute.

All this, however, still did not satisfy the pious priests. By threats of eternal hell-punishment they formally extorted ever larger donations, so that Charlemagne as late as 811, in the Capitulary of Aachen, reproached them with this and, in addition, with “seducing people to perjury and false witness in order to increase your wealth (that of the bishops and abbots)”. Unlawful donations were obtained by trickery, in the confidence that the Church, apart from her privileged jurisdiction, still possessed means enough to make a fool of justice. In the 6th and 7th centuries there was scarcely a Gallic council that did not threaten with ecclesiastical anathema any and every challenge to donations to the Church. Even formally invalid donations were to be converted by this method into valid ones, and the private debts of individual clerics protected from collection. “Truly contemptible are the means we see employed, time and again, to reawaken the desire to make donations. When the descriptions of heavenly bliss and infernal torments no longer sufficed, relics were brought from distant parts, translations held, and new churches built; in the 9th century this was a regular branch of business.” (Roth, p. 254.) “When the envoys of the Monastery of Saint Médard of Soissons, after great exertions in Rome, had begged the body of Saint Sebastian and additionally stolen that of Saint Gregory, and both had been laid to rest in the monastery, so many people flocked to the new saints that the region was as though sown with locusts, and those seeking help were healed not singly but in whole herds. The result was that the monks measured the money in bushels, of which they counted 85, and that their store of gold amounted to 900 pounds.” (p. 255.) Fraud, legerdemain, apparitions of the dead, especially of saints, served to swindle riches for the Church, and finally, and principally—the forging of documents. This, to let Roth speak again, was “carried on by many clerics on a massive scale … this business began very early … The extent to which this trade was practised is evident from the large number of forged documents which our collections contain. Among the 360 Merovingian diplomas in Bréquigny, approximately 130 are decidedly false. … The false testament of Remigius was already used by Hincmar of Rheims to procure for his church a series of possessions of which the genuine one says nothing, although the latter was never lost and Hincmar knew very well that the former was spurious.” Even Pope John VIII attempted to acquire possession of the Monastery of Saint Denis near Paris by means of a deed which he knew to be false. (Roth, pp. 256 ff.)

Thus we cannot be surprised if the landed property of the Church, scraped together through donation, extortion, trickery, fraud, forgery, and other penitentiary industries, assumed quite colossal proportions within a few centuries. The Monastery of Saint-Germain-des-Prés, now within the bounds of Paris, possessed at the beginning of the 9th century landed property of 8,000 mansi, or hides, whose area Guérard calculates at 429,987 hectares, with an annual yield of 1 million francs = 800,000 marks. If we take as a basis the same average of 54 hectares of area and 125 francs = 100 marks of income per hide, then at about the same time the monasteries of Saint Denis, Luxeuil, and Saint Martin of Tours each had landed property of some 15,000 mansi, an area of 810,000 hectares, and an income of 1½ million marks. And this was after the confiscation of Church property by Pepin the Short!—As for the total ecclesiastical property in Gaul at the end of the 7th century, Roth (p. 249) estimates it at rather more than less than one-third of the total area.

These enormous masses of estates were cultivated partly by unfree, but also partly by free, dependents of the Church. Of the unfree, the slaves (servi) were originally subject to unlimited services to their lords, since they were not legal persons; but here, too, a customary measure of dues and services soon appears to have established itself for settled slaves. The services of the two other unfree classes, the coloni and liti (as to whose legal distinction for that period no information exists), were on the other hand fixed and consisted of certain manual and team services, as well as a definite share of the yield of the estate. These were relations of dependence of long-standing tradition. By contrast, it was something new for Germans that free men should sit on land other than common land or their own. To be sure, the Germans in Gaul and generally in the territory subject to Roman law often enough found free Romans as tenants; but that they themselves should not need to become tenants, but could sit on their own land, was ensured by the land settlement. Before, therefore, free Franks could become the dependents of anyone at all, they had first to lose the alod which they had received at the land settlement in some way or other; a separate class of landless free Franks had first to form.

This class formed itself through the incipient concentration of landed property, through the same causes from which that concentration arose: on the one hand, through the civil wars and confiscations; on the other, through the transfers of land to the Church, largely occasioned by the pressure of circumstances, by the desire for security. And the Church soon found yet another special means to promote such transfers, by not merely leaving the donor his estate in usufruct against tribute, but also granting him in addition a piece of church land in tribute. These donations, namely, occurred in a double form. Either the donor reserved the usufruct of the estate for life, so that it passed into the Church’s ownership only after his death (donatio post obitum); in this case it was customary, and was later expressly laid down in the capitularies of the kings, that the donor should receive from the Church twice the amount of the donated estate in tribute. Or the donation became effective at once (cessio a die praesente), and then the donor received three times the amount in church land, together with his own estate, in tribute by means of a so-called precaria, a document issued by the Church, which transferred these parcels of land to him, mostly for life, but sometimes also for a longer or shorter period. Once the class of landless freemen had been created, many of them also entered into such a relationship; the precariae granted to them seem at first to have been issued mostly for five years, but here, too, they soon became lifelong.

It can hardly be doubted that already in the Merovingian period quite similar relations developed on the estates of the secular magnates as on church lands, so that there, too, alongside the unfree, free dependents were settled in tribute. They must even have been very numerous by the time of Charles Martel, because otherwise the revolution in landed property relations begun by him and completed by his son and grandson would remain inexplicable at least from one side.

This revolution is based in substance on two new institutions. First, in order to bind the magnates of the realm to the crown, from now on the crown lands were as a rule no longer given to them but merely granted for life, as a “beneficium”, and this under specific conditions which had to be observed on pain of forfeiture. In this way, they themselves became dependents of the crown. And second, in order to ensure the provision of the free dependents of the magnates for military service, part of the official powers of the Gau counts over the freemen settled on their estates were transferred to these magnates, who were appointed as their “seniors”. Of these two changes we have here, for the present, to consider only the first.

In suppressing the rebellious petty “tyrants”, Charles—the sources are lacking—will, in accordance with old custom, have confiscated their landed property, but will have regranted it to them, wholly or in part, as a beneficium, insofar as he afterwards restored them to office and dignities. With the church property of recalcitrant bishops he did not yet dare to proceed thus; he deposed them and gave their posts to men devoted to him, some of whom, to be sure, had nothing of the cleric but the tonsure (sola tonsura clericus). These new bishops and abbots now began, at his command, to transfer large tracts of church land to laymen as precariae; this had not been without precedent earlier, but now it was done on a massive scale. His son Pepin went considerably further. The Church was declining, the clergy despised, the pope hard pressed by the Lombards and solely dependent upon the help of Pepin. Pepin aided the pope, favoured the extension of his ecclesiastical dominion, held the stirrup for him. But he made himself paid by incorporating by far the greatest part of the church property into the crown lands and leaving the bishops and monasteries only what was necessary for their maintenance. This first secularisation on a large scale was accepted by the Church without resistance; the Synod of Lestines confirmed it, albeit with a restrictive clause which was, however, never observed. This enormous mass of property restored the exhausted crown lands to a respectable footing and served in great part for further grants, which in fact soon assumed the form of ordinary benefices.

Let us insert here that the Church very soon knew how to recover from this blow. Scarcely had the settlement with Pepin been effected when the good men of God began their old practices again. Donations rained down once more from all sides, the little free peasants were continually in the same dire straits between hammer and anvil as they had been for 200 years; under Charlemagne and his successors it went even worse with them, and many betook themselves, with house and homestead, under the shelter of the crozier. The kings returned part of the plunder to favoured monasteries, made gifts of enormous tracts of crown land to others, especially in Germany; under Louis the Pious the blessed times of Guntram seemed to have come again for the Church. The archives of the monasteries are particularly rich in donations from the ninth century.

The benefice, this new institution which we now have to investigate more closely, was not yet the later fief, but was certainly its germ. From the outset it was granted for the joint lifetime of both the grantor and the recipient. If either one or the other died, it fell back to the owner or his heirs. To renew the previous relationship, a new grant to the recipient or his heirs had to take place. The benefice was thus, in the later phraseology, subject to throne-fall as well as to reversion. Throne-fall soon fell into disuse; the great beneficiaries became precisely more powerful than the king. Reversion, from an early date, not infrequently entailed the further grant of the estate to the heirs of the previous beneficiary. An estate Patriciacum (Percy) near Autun, granted by Charles Martel to Hildebrannus as a benefice, remained in the family from father to son through four generations, until the king gave it to the brother of the fourth beneficiary as full property in 839. Other examples are not rare from the middle of the 8th century onwards.

The benefice could be withdrawn by the grantor in all cases in which forfeiture of property was incurred at all. And such cases were more than abundant even under the Carolingians. The uprisings in Alemannia under Pepin the Short, the conspiracy of the Thuringians, the repeated uprisings of the Saxons led to ever new confiscations, whether of free peasants’ land, or of the estates and benefices of the great. The same was the case, in defiance of all contractual provisions to the contrary, during the civil wars under Louis the Pious and his sons. Certain non-political crimes also entailed confiscation.

Furthermore, benefices could be withdrawn by the crown if the beneficiary neglected his general duties as a subject, e.g. did not deliver up the robber from the immunity, did not bring his harness on the campaign, failed to respect royal letters, etc.

Then, however, benefices were granted under specific conditions, the breach of which brought withdrawal in its wake, by which, self-evidently, the remainder of the beneficiary’s property was not affected. Thus, for example, if former church lands had been granted and the beneficiary neglected to pay the dues owed to the church (nonae et decimae) with which they were encumbered. Thus if he allowed the estate to fall into decay, in which case a one-year warning period was usually set so that the beneficiary could protect himself from the otherwise ensuing confiscation by making improvements, etc. Furthermore, the grant of the estate could also be tied to specific services, and in fact it became so more and more to the extent that the benefice developed further into the actual fief. But originally this was by no means necessary; least of all as far as military service is concerned; a great many benefices were granted to lower clergy, to monks, to ecclesiastical and lay women.

Finally, it is in no way excluded that the crown initially also granted lands revocably or for a fixed term, i.e. as precaria. Isolated reports and the practice of the Church make it probable. In any case, however, this soon ceased, since granting on beneficiary terms became general in the 9th century.

For the Church – and we must assume the same of the great landowners and beneficiaries – the Church, which previously had mostly transferred lands to its free dependent tenants only as precaria for a term, had to follow the impulse given by the crown. Not only did it begin to grant benefices as well, but this mode of granting became so predominant that the already existing precaria became lifelong and imperceptibly assumed the nature of the benefice, until in the ninth century the former merged almost entirely into the latter. In the latter half of the 9th century the beneficiaries of the Church, and likewise those of the secular magnates, must already have assumed an important position in the state; a number of them must have been people of considerable property, founders of the later lower nobility. Otherwise Charles the Bald would hardly have intervened so vigorously on behalf of those from whom Hincmar of Laon had taken their benefices without cause.

We see that the benefice already has many features that reappear in the developed fief. Throne-fall as well as reversion are common to both. Like the fief, the benefice is subject to withdrawal only under specific conditions. In the social hierarchy created by the benefices, which descends from the crown through the great beneficiaries – the predecessors of the imperial princes – to the middling beneficiaries – the later nobility – and from these to free and unfree peasants, for the most part living in the mark association, we see the foundation of the later closed feudal hierarchy. While the later fief is under all circumstances a service-tenement and obliges to military service for the lord of the fief, the latter is admittedly not yet the case with the benefice, and the former is by no means necessary. But the tendency of the benefice to become a service-tenement is already unmistakably present and gains more and more scope in the course of the 9th century; and to the same extent as it unfolds freely, the benefice develops into the fief.

In this development, however, a second lever cooperates: the transformation which the Gau and army constitution underwent, first under the influence of large landed property and later under that of the great benefices, into which the former large landed property had more and more been converted as a result of the incessant civil wars and the confiscations and new grants connected therewith.

It is understandable that in this chapter we speak only of the benefice in its pure, classical form, in which it was, to be sure, only a vanishing form, one that did not even appear everywhere simultaneously. But such historical forms of appearance of economic relations are understood only when one grasps them in this purity of theirs, and to have peeled this classical form of the benefice out of all the confused accretions is one of the principal merits of Roth.

Gau and Army Constitution.

The upheaval in the condition of landed property just depicted could not remain without influence on the old constitution. It brought about equally significant changes in the latter, and these in turn reacted on the relations of landed property. We leave aside for the time being the transformation of the general state constitution, and confine ourselves here to the influence of the new economic situation on the still persisting remnants of the old popular constitution in Gau and army.

As early as under the Merovingians we frequently find counts and dukes acting as administrators of crown land. It is not until the 9th century, however, that we find indisputably that certain crown lands were connected with the office of count in such a way that the count of the time received their income. The earlier honorary office had been transformed into one salaried by endowment. Alongside this we also find, as a matter of course under the contemporary conditions, counts in possession of royal benefices assigned to them personally. Thus the count became a powerful landowner within his county.

To begin with it is clear that the authority of the count was bound to suffer from the rise of great landowners below and beside him. People who under the Merovingians and the first Carolingians often enough scoffed at the king’s commands were bound to show even less respect for the count’s order. Their free dependent tenants, trusting in the protection of powerful lords, just as frequently neglected to obey the count’s summons to court, or his call to the host. This was precisely one of the causes that brought about the introduction of granting as a benefice instead of as allod, and the later gradual conversion of the greater part of the formerly free large landed property into benefices.

This alone still did not secure the bringing in of the freemen settled on the estates of the magnates for state duties. A further change had to occur. The king found himself compelled to make the great landowners responsible for the appearance of their free dependent tenants in court, in the army, and for other customary state services in the same way as the count had previously answered for all free inhabitants of his county. And this could only happen by the king transferring to the magnates a part of the count’s official prerogatives over their dependent tenants. The landlord or beneficiary had to present his people in court; they thus had to be summoned through his mediation. He had to lead them to the host; their summons thus had to take place through him; he had, in order to be able to answer for them continuously, to have command over them and the right of military discipline. But it was, and remained, king’s service on the part of the dependent tenants; the refractory were punished not by the landowner but by the royal count; the penalty fell to the royal fisc.

This innovation, too, goes back to Charles Martel. At least it is only from his time that we find the custom of the great ecclesiastical dignitaries themselves taking the field, which according to Roth can only be explained by the fact that Charles had the bishops join the host at the head of their dependent tenants in order to assure himself of the appearance of the latter. Undoubtedly the same happened with the secular magnates and their dependent tenants. Under Charles the Great the new institution appears already firmly established and universally carried through.

Herewith, however, an essential change had come about in the political position of the free dependent tenants as well. They, who previously were legally on an equal footing with their landlord, however much they may have been dependent on him economically, now became legally his subordinates as well. The economic subjugation received political sanction. The landlord becomes senior, seigneur, the dependent tenants become his homines; the “lord” becomes the superior of the “man.” The legal equality of the freemen is gone; the lowest “man,” whose full freedom has already been severely impaired by the loss of his inherited land, draws a step closer to the unfree. So much the more does the new “lord” raise himself above the level of the old common freedom. The economically already produced foundation of the new aristocracy is acknowledged by the state, becomes one of the regularly cooperating driving wheels of the state machine.

Alongside these homines consisting of free dependent tenants, however, there was yet another kind. These were the impoverished freemen who had voluntarily entered into a relationship of service or following with the great. The Merovingians had had their retinue in the antrustions; the magnates of that time will likewise not have been without a following. Under the Carolingians the retinue of the king are called vassi, vasalli or gasindi, expressions which in the oldest folk-laws are still used for an unfree person, but which have now already assumed the meaning of a normally free retainer. The same designations apply to the retainers of the magnates, who now appear quite generally and become an ever more numerous and important element in society and the state.

How the great men came by such retainers is shown by the old contract formulas. In one of these (formulae Sirmondicae 44) it says, for example:[21] “Forasmuch as it is known to all men that I have not wherewithal to feed and clothe myself, I pray by your (the lord’s) piety that I may place myself under your protection (*mundoburdum*—that is, guardianship) and commend myself to you. In such wise… that you shall be bound to help me with food and clothing in proportion as I serve you and shall merit it; and I, so long as I live, shall be bound to render you service and obedience as a free man (*ingenuili ordine*); nor shall I have the power to withdraw from your power and protective authority during my lifetime, but shall remain all my life under your power and protection.” This formula gives full information on the origin and nature of the simple relation of retainership, stripped of all extraneous admixture, and all the more so because it presents the extreme case of a poor devil who has come down completely in the world. Entry into the lord’s body of retainers took place by a free agreement of both parties—free in the sense of Roman and modern jurisprudence, often [366] Franconian Time enough like the entry of a present-day worker into the service of a manufacturer. The “man” commended himself to the lord, and the lord accepted his commendation. This consisted in a handclasp and oath of fealty. The agreement was for life and was dissolved only by the death of one of the contracting parties. The servant was bound to perform all contractual services compatible with the position of a freeman that his lord might impose on him. In return he was maintained by the latter and rewarded according to discretion. An assignment of land was by no means necessarily connected with this, and indeed did not take place in all cases.

[22] This relationship was not only tolerated under the Carolingians, especially since Charlemagne, but directly favoured and at last, as it seems, made obligatory for all common freemen and regulated by the state through a capitulary of 847. Thus the servant could only unilaterally dissolve the relationship with his lord if the latter sought to kill him, beat him with a stick, dishonour his wife or daughter, or take his inheritance from him (Capitulary of 813). And the servant was bound to the lord as soon as he had received from him the value of a solidus; whence it again clearly emerges how little the relationship of vassalage was necessarily tied to the grant of land at that time. The same provisions are repeated in a capitulary of 816, with the addition that the servant is released if his lord seeks unlawfully to bring him into the status of unfreedom or could provide him the promised protection but does not do so.

In relation to the state, the retainer-lord now acquired the same rights and obligations with regard to his retainers as the landowner or *beneficiarius* with regard to his dependants. They remained bound to perform service to the king, only here, too, the retainer-lord now interposed himself between the king and his counts. He presented the vassals before the court, he summoned them, led them in war and maintained discipline among them, he was answerable for them and their prescribed equipment. Thereby, however, the retainer-lord acquired a certain punitive authority over his subordinates, and this formed the starting-point of the later developing jurisdiction of the feudal lord over his vassals.

[23] In these two further institutions—in the elaboration of the system of retainership and in the transfer of countly, i.e. state, official authority to the landowner, royal *beneficiarius* and retainer-lord over his subordinates, now soon all combined under the names *vassi*, *vasalli*, *homines*—dependants as well as landless retainers—in this state confirmation and reinforcement of the lord’s actual power over the vassals, we see the seed of feudalism given in the benefices already developing considerably further. The hierarchy of estates from the [367] Friedrich Engels king downwards through the great *beneficiarii* to their free dependants and finally down to the unfree, becomes a recognised element of the state order, co-operating in an official capacity. The state recognises that it cannot exist without their help. How this help was actually rendered will, of course, be shown.

The distinction between retainers and dependants is only important for the beginning, in order to demonstrate the dual origin of the dependence of freemen. Very soon both types of vassals coalesce inseparably, in name as well as in fact. The great *beneficiarii* more and more adopted the practice of commending themselves to the king, thus becoming, alongside his *beneficiarii*, his vassals. The kings found it in their own interest to have the oath of fealty sworn to them personally by the great men, bishops, abbots, counts and vassals (Annales Bertiniani 837 and frequently in the 9th century), by which process the difference between the general subject’s oath and the special vassal’s oath was bound to become blurred. Thus gradually all the great men transformed themselves into royal vassals. Herewith, however, [24] the slow development of the great landowners into a distinct estate, into an aristocracy, was recognised by the state, incorporated into the state order, and made into one of its officially functioning levers.

In the same way, the retainer of the individual great landowner gradually merges into the dependant. Apart from direct provision at the lord’s court, which could take place only for a small number of heads, there remained no other means of securing retainers than settling them on land and soil, transferring land to them as a benefice. A numerous combat-ready retinue, the main condition of the existence of the great ones in that period of eternal struggles, was therefore only to be obtained by granting land to the vassals. Hence the landless servants of the lord’s court gradually disappear before the mass of those settled on the lord’s land.

But the more this new element pushed itself into the old constitution, the more this was bound to be shaken. The old immediate exercise of state power by king and counts more and more gave way to an indirect one; between the common freemen and the state stepped the lord, to whom the former were bound by personal fealty on an ever-growing scale. The most effective driving part of the state machine, the count, had to recede more and more into the background, and actually did so. Charlemagne proceeded here as he was accustomed to proceed everywhere. At first he favoured, as we have seen, the predominance of the vassal relationship, until the independ-[25]ent small freemen had almost vanished; when then the weakening of his power brought about by this came to light, he tried to help it to its feet again by state intervention. That might succeed in many [368] Franconian Time cases under so energetic and feared a ruler; under his weak successors, the power of the facts created with his help irresistibly broke its way.

Charles’s favourite means was the dispatch of royal envoys (*missi dominici*) with extraordinary plenipotentiary powers. Where the ordinary royal official, the count, could not check the spreading disorder, a special envoy was to do this. (This to be historically substantiated and developed further.)

Now, however, there was yet another means, and this consisted in placing the count in such a position that he was at least equal to the great men of his county in material means of power. This was only possible if the count likewise entered the ranks of the great landowners, which again could happen in two ways. Certain landed properties could be attached in the individual districts as an endowment to the count’s office, so that the respective count administered them *ex officio* and drew their revenues. Many examples of this are found especially in deeds, and indeed from the end of the 8th century; from the 9th on this relationship is quite usual. Such endowments naturally stem mostly from the royal fisc, just as we already find, in Merovingian times, counts and dukes frequently as administrators of the royal fisc lands lying in their territory.

[26] Remarkably, there are also many examples (even a formulary for it) where bishops endow the count’s office from the church property, naturally, given the inalienability of church property, in some form of benefice. The liberality of the church is too well known for any other reason to be acceptable here than bitter necessity. Under the growing pressure of the neighbouring secular great men, the only recourse left to the church was alliance with the remnants of the state power.

These perquisites connected with the count’s posts (*res comitatus, pertinentiae comitatus*) are initially still sharply distinguished from the benefices that were personally conferred on the particular count. These, too, were usually and richly granted; so that, endowment and benefices taken together, the count’s offices, originally posts of honour, now became very lucrative positions and, from Louis the Pious onwards, were given away just like other royal largesses to men whom one wished to win over or of whom one wished to make sure. Thus it says of Louis the Stammerer that *quos potuit conciliavit, dans eis comitatus, et abbatias ac villas* (Annales Bertiniani 877). The designation *honor*, which earlier denoted the office with reference to the rights of honour attaching to it, acquires in the course of the 9th century quite the same meaning as *beneficium*. And herewith a substantial change necessarily took place in the [369] Friedrich Engels

370

Frankish Period

Character of the Countship, rightly emphasised by Roth (p. 408) \[V, 27\]. Originally the seniorate, insofar as it acquired a public character, was modelled on the countship and equipped with countly powers. Now—in the second half of the 9th century—the seniorate had spread so universally that it threatened to overrun the countship, and the latter could only maintain its position of power by itself assuming more and more the character of a seniorate. The counts increasingly and not without success usurped the position of a senior towards the inhabitants of their Gau (pagenses), with regard to both their private and their public relations. Just like the other “lords” with the lesser people in their neighbourhood, the counts too sought, by fair means or foul, to bring the less well-off free inhabitants of the Gau to submit to them as vassals. This succeeded all the more easily as the mere fact that the counts were able to misuse their official authority in this way is the best proof of how little protection the still remaining remnant of the common freemen could expect from the royal power and its organs. Exposed on all sides to violent oppression, the smaller freemen had to be glad to find some protector at all, even if it meant surrendering their allod and receiving it back as a mere benefice. As early as the Capitulary of 811 Charlemagne complains that bishops, abbots, counts, judges, and centenarii, by constant legal chicanery or by repeatedly calling out the army, so hound the lesser people that they transfer or sell their allod to them; that the poor clamour about the robbery committed against their property, etc. In this way, by the end of the 9th century the greater part of free property in Gaul had already passed into the hands of the Church, the counts, and other magnates (Hincmar, Annales Remenses, 869), and somewhat later in some provinces there was no longer any free landed property of small freemen at all. (Maurer, Einleitung, 212) Now as the benefices gradually became hereditary, given the growing power of the beneficiaries and the declining power of the crown, the countships too became hereditary by custom. If in the mass of royal beneficiaries we saw the rudiments of the later nobility, here we have the germ of the territorial sovereignty of the later territorial lords who emerged from the gau counts.

While the social and political order was thus completely transformed, the old military constitution, based on military service—both a right and a duty—of all freemen, remained outwardly the same; only that, where the new relations of dependence existed, the senior interposed himself between his vassals and the count. But the common freemen were less and less able, from year to year, to bear the burden of military service. This consisted not only in personal service; the man called up also had to equip himself and maintain himself at his own expense for the first six months, until at last the incessant wars of Charlemagne brought matters to a head. The burden became so unbearable that, in order to escape it, the small freemen in masses preferred to transfer not only the remainder of their possessions, but their own persons and those of their descendants to the magnates, ||\[28\]j and especially to the Church. Thus Charlemagne had reduced the free, warlike Franks to the point where they preferred to become bondsmen and serfs merely to avoid going to war. This was the result of Charlemagne insisting on carrying through, and pushing to the utmost extreme, a military constitution based on the universal and equal landholding of all freemen, at a time when the great mass of freemen had already wholly or largely lost their landed property.

Yet the facts were stronger than Charlemagne’s obstinacy and ambition. The old military constitution could no longer be maintained. To equip and maintain the army at state expense was even less possible in that period of an almost moneyless and tradeless natural economy. Charlemagne was therefore compelled to restrict the duty of service in such a way that the man’s equipment and maintenance remained feasible. This was done in the Capitulary of Aachen of 807, when the wars were limited to frontier struggles and the existence of the empire as a whole seemed assured. Above all, every royal benefice-holder was required to appear without exception. Then, anyone possessing 12 hides (mansi) was to appear in armour, thus probably also on horseback (the word caballarius, knight, occurs in the same capitulary). Possessors of 3 to 5 hides were liable. From every two possessors of 2 hides each, from every three with one hide, from every six with half a hide, one man had to be furnished and equipped by the others. Of freemen entirely landless but possessing movable property worth 5 solidi, the sixth man was also to march out and receive a money payment of one solidus from each of the other five. The obligation to march out of the various parts of the country, which fully applied in these neighbouring wars, is also restricted for more distant wars according to the distance to one half to one sixth of the men.

Charlemagne was obviously here trying to adapt the old constitution to the changed economic conditions of those liable for service, to save what could still be saved. But even this concession did not help; soon after, ||\[29\]| he was compelled in the Capitulare de exercitu promovendo to allow new exemptions. This capitulary, usually dated earlier than that of Aachen, is undoubtedly several years later than the latter according to its entire content. It raises the number of hides for which one man had to be furnished from three to four; the possessors of half hides and the landless appear as exempt from service, and for benefice-holders, too, the obligation to provide men is limited to one man for every four hides. Under Charlemagne’s successors, the minimum number of hides that furnished one man seems even to have been raised to five.

It is remarkable that the provision of the armoured twelve-hide men seems to have encountered the greatest difficulties. At least the order that they were to appear in armour is repeated countless times in the capitularies.

Thus the common freemen disappeared more and more. While their gradual separation from the land had driven one part into the vassalage of the new great landowners, the fear of direct ruin through military service drove the other part straight into serfdom. How rapidly this submission to bondage proceeded is attested by the Polyptych (land register) of the monastery of Saint-Germain-des-Prés, which at that time still lay outside Paris. It was compiled by Abbot Irmino at the beginning of the 9th century and lists among the dependents of the monastery: 2,080 families of coloni, 35 of liti, 220 of slaves (servi), and only eight free families. Yet the word colonus at that period definitely designated an unfree person in Gaul. The marriage of a free woman with a colonus or slave rendered her as defiled (deturpatam) subject to the lord (Capitulary of 817). Louis the Pious orders that a colonus vel servus (of a monastery at Poitiers) shall return to natural servitude whether he will or no (velit nolit redeat). They received blows (Capitularies of 853, 861, 864, 873) and were sometimes manumitted (see Guérard, “Irmino”). And these serf peasants were by no means Romans; according to Jacob Grimm’s own testimony (Geschichte der deutschen Sprache, I, p. \[537\]), who examined the names, they were “almost exclusively Frankish, far outnumbering the small number of Roman ones.”

This enormous increase of the unfree population again shifted the class relations ||\[30\]| of Frankish society. Alongside the great landowners, who were then rapidly developing into a distinct estate, and alongside their free vassals, there now appeared a class of unfree persons that increasingly absorbed the remainder of the common freemen. But these unfree persons had partly themselves still been free, partly were children of freemen; those who had lived in hereditary bondage for three generations and more were by far the minority. Nor were they for the most part Saxon, Wendish, etc., prisoners of war imported from outside; on the contrary, the majority were native Franks and Romans. With such people, when they moreover were beginning to make up the mass of the population, it was not as easy to deal as with hereditary or foreign serfs. Bondage was still unfamiliar to them; the blows which even the colonus received (Capitularies 853, 861, 873) were still felt as an indignity, not as something to be taken for granted. Hence the numerous conspiracies and uprisings of the unfree, and even of the peasant vassals. Charlemagne himself forcibly put down a rising of the dependents of the bishopric of Reims. In the Capitulary of 821, Louis the Pious speaks of conspiracies of the slaves (servorum) in Flanders and Menapiscus (on the upper Lys). In 848 and 866, risings of the service men (homines) of the bishopric of Mainz had to be suppressed. The orders to suppress such conspiracies recur in the capitularies from 779 onward. The rising of the Stellinga in Saxony must also belong here. It was obviously a consequence of this menacing attitude of the unfree masses when, from the end of the 8th and the beginning of the 9th century, the services of the unfree, even of the settled slaves, were increasingly fixed at a definite, unsurpassable measure, and Charlemagne prescribes this in his capitularies.

This, then, was the price at which Charlemagne purchased his new Roman empire: the destruction of the estate of the common freemen, which at the time of the conquest of Gaul had embraced the entire Frankish people; the division of the people into great landowners, vassals, and serfs. But with the common freemen the old military constitution fell, and with both the kingship fell. Charlemagne had destroyed the sole foundation of his own rule. He himself still held out; but under his successors what was in reality the work of his hands came to light.

\[\[31\]j (Note)

The Frankish Dialect.

Philologists have played strange tricks with this dialect. Whereas Grimm let it be swallowed up in French and High German, more recent scholars give it an extension reaching from Dunkirk and Amsterdam to the Unstrut, Saale, and Rezat, if not as far as the Danube and, through colonisation, to the Riesengebirge. While even a philologist like Moritz Heyne constructs an Old Low Franconian language out of a manuscript of the Heliand made at Werden—which is almost pure Old Saxon with a very slight Frankish tinge—Braune simply assigns all genuinely Low Franconian dialects now to Saxon, now to Netherlandish. And finally, Arnold restricts the area of conquest of the Ripuarians to the district north of the watershed of the Ahr and Moselle and leaves everything to the south and south-west to have been settled first by Alamanni, later exclusively by Chatti (whom he also counts as Franks), so that Alemannic-Chattish was spoken there.

Let us first reduce the Frankish language area to its real boundaries. Thuringia, Hesse, and Main Franconia have absolutely no other claim to be counted as part of it than that they were included under Francia in Carolingian times. The language spoken east of the Spessart, Vogelsberg, and the Kahler Asten is everything but Frankish. Hesse and Thuringia have their own independent dialects, just as they are inhabited by independent tribes; in Main Franconia a mixture of Slav, Thuringian, and Hessian population has been intermingled with Bavarian and Frankish elements and has developed its own distinct dialect. Only if one applies the degree to which the High German sound-shift has penetrated the dialects as the main means of differentiation can one assign these three branches of language to Frankish. But it is precisely this procedure, as we shall see, that has caused all the confusion in the assessment of Frankish language by non-Franks.

Let us begin with the oldest monuments, and first put Moritz Heyne’s*[1] so-called Old Low Frankish in its proper light. The so-called Cotton manuscript of the *Heliand*, produced in Werden and now in Oxford, is said to be Old Low Frankish because it was produced in the monastery of Werden, still on Frankish soil, but hard by the Saxon border. Here the old tribal border is still today the border between Berg and Mark; of the abbeys lying in between, Werden belongs to Franconia, Essen to Saxony. Werden is bounded to the east and north, in the very closest proximity, by indisputably Saxon settlements; in the plain between the Ruhr and Lippe, Saxon language penetrates in places almost to the Rhine. The circumstance that a Saxon work was copied in Werden, and indeed manifestly by a Frank, and that every now and then Frankish word-forms flowed from this Frank’s pen, is far from sufficient to declare the language of the copy Frankish. – In addition to the Cotton *Heliand*, Heyne considers as Low Frankish some Werden fragments showing the same character and the remains of a Psalter translation which, according to him, originated in the Aachen area, but which Kern (*Glossen in der Lex Salica*) briefly declares to be Low Dutch. In fact, it has on the one hand quite Low Dutch forms, alongside them however also genuinely Rhenish Frankish ones, and even traces of High German sound-shift. It evidently originated on the border between Low Dutch and Rhenish Frankish, roughly between Aachen and Maastricht. Its language is considerably later than that of the two *Heliand*s.

The Cotton *Heliand* alone, however, suffices to establish beyond doubt, from the few Frankish forms occurring in it, some principal differences between Frankish and Saxon.

I. In all Ingvaeonic dialects the three persons of the plural present indicative end alike, namely in a dental with a preceding vowel; Old Saxon in *-d*, Anglo-Saxon in *-dh*, Old Frisian in *-th* (which probably also stands for *-dh*). Thus in Old Saxon we have *hebbiad* we have, you have, they have; likewise from *fallan*, *gewinnan*, all three persons similarly *fallad*, *winnad*. It is the third person that has taken over all three, but it should be noted, with the specifically Ingvaeonic ejection of *n* before the *-d* or *-dh*, likewise common to all three dialects named. Of all living dialects, this peculiarity has been preserved only by Westphalian; there one still says today *wi*, *ji*, *se hebbed*, etc. The other Saxon dialects, like West Frisian, no longer know it; they distinguish the three persons.

The west-Rhenish Psalms have, like Middle High German, for the 1st person plural *-n*, 2nd *-t*, 3rd *-nt*. The Cotton *Heliand*, on the other hand, alongside the Saxon ones, occasionally has forms of quite another kind: *tholönd* they suffer, *gornönd* you lament, and as imperative *marient* proclaim, *seggient* say, where Saxon demands *tholöd*, *gornöd*, *märiad*, *seggiad*. These forms are not only Frankish, they are even genuine Werden, Berg local dialect to this day. In the Berg dialect we likewise make all three plural persons of the present the same, but not in Saxon on *-d*, but in Frankish on *-nt*. As against Markish *wi hebbed*, right on the border one says *wi hant*, and analogous to the imperative *seggient* given above, one says *seient ens*, say once (i.e., do say). Braune and others, on the basis of the simple observation that here in Berg the three persons are made alike, have summarily declared the whole Berg mountainous country to be Saxon. The rule certainly has penetrated over from Saxony; unfortunately, however, it is executed in Frankish fashion, thus proving the opposite of what it is supposed to prove.

The ejection of *n* before dentals in the Ingvaeonic dialects is not confined to this case; in Old Frisian it is less widespread, in Old Saxon and Anglo-Saxon on the other hand quite extensive: *mudh* mouth, *kudh* known, *us* us, *odhar* another. The Frankish copyist of the *Heliand* in Werden writes instead of *odhar* twice the Frankish form *andar*. The Werden rent-rolls alternate between the Frankish name-forms *Reinswind*, *Meginswind*, and the Saxon *Reinswid* and *Meginswid*. In the left-Rhenish Psalms, on the other hand, it is everywhere *munt*, *hunt*, *uns*, only once do the so-called Lipsius glosses (extracted from the lost manuscript of these Psalms) have *farkutha* abominabiles instead of *farkuntha*.

The Old Salian monuments have likewise preserved the *n* everywhere in the names *Gund*, *Segenand*, *Chlodosindis*, *Ansbertus*, etc., which is of no account. The modern Frankish dialects have the *n* everywhere (sole exception in Berg the form *os*, us).

II. The linguistic monuments from which the so-called Old Saxon grammar is usually constructed all belong to south-western Westphalia: Münster, Freckenhorst, Essen. The language of these monuments shows some essential deviations not only from the general Ingvaeonic forms, but also from those which have come down to us in proper names from Angria and Eastphalia as genuinely Old Saxon; on the other hand, they agree remarkably with Frankish and Old High German. The most recent grammarian of the dialect, Cosijn, therefore even calls it straightforwardly Old West Saxon.

Since we are almost exclusively reliant on proper names in Latin documents for this investigation, the demonstrable formal differences between West and East Saxon can be only few in number; they are confined to two, but very decisive, cases.

1) Anglo-Saxon and Old Frisian have the genitive plural of all declensions in *-a*. Old West Saxon, Old Frankish and Old High German, in contrast, in *-o*. Now, what is the correct Old Saxon form? Did this dialect here really abandon the Ingvaeonic rule?

The documents from Angria and Eastphalia provide the answer. In *Stedieraburg*, *Horsadal*, *Winethahiisen*, *Edingahüsun*, *Magathaburg* and many other names, the first element of the compound is in the genitive plural and has *-a*. Even in Westphalia the *-a* has not yet entirely vanished; the Freckenhorst Roll once has *Aningera lö* and *Wernera holthüson*, and the *-a* in Osnabrück is likewise an old genitive plural.

2) Likewise, the weak masculine in Frankish ends in *-o*, as in Old High German, as against Gothic-Ingvaeonic *-a*. For Old West Saxon, *-o* is likewise the fixed rule; thus again a departure from Ingvaeonic usage. This, however, by no means holds for Old Saxon as a whole. Not even in Westphalia did *-o* apply without exception; the Freckenhorst Roll already has alongside *-o* a whole series of names in *-a* (*Siboda*, *Uffa*, *Asica*, *Hassa*, *Wenda*, etc.); Paderborn monuments in Wigand yield almost always *-a*, only quite exceptionally *-o*; in Eastphalian documents *-a* prevails almost exclusively; so that already J. Grimm (*Geschichte der deutschen Sprache*) arrives at the conclusion that it cannot be denied that *-a* and *-an* (in oblique cases) was the original Saxon form common to all parts of the people. The advance of *-o* for *-a* was also not confined to Westphalia. At the beginning of the 15th century, the East Frisian men’s names in the chronicles etc. have almost regularly *-o*: *Fokko*, *Occo*, *Enno*, *Smelo*, etc., as against earlier *-a* still preserved in isolated West Frisian instances.

It may therefore be taken as established that both deviations of West Saxon from the Ingvaeonic rule are not originally Saxon, but are due to foreign influence. This influence is explained very simply by the fact that West Saxony was formerly Frankish territory. Only after the withdrawal of the main mass of the Franks did the Saxons advance gradually over the Osning and Egge mountains as far as the line that still today separates Mark and Sauerland from Berg and Siegerland. The influence of the Franks who remained behind, now fused with the Saxons, shows itself in those two *-o* instead of *-a*; it is also still unmistakable in the present-day dialects.

III. A peculiarity of Rhenish Frankish extending from the Ruhr to the Moselle is the ending of the 1st person present indicative in *-n*, which is best preserved in the case where a vowel follows: *dat don ek* that do I, *ek han* I have (Bergish). This verbal form holds for the entire Lower Rhine and the Moselle at least as far as the Lotharingian border: *don*, *han*. The same peculiarity is already found in the left-Rhenish Psalms: *biddon* I pray, *wirthon* I become, although not consistently. This *-n* is absent from the Salian dialect; in the oldest document there already we have: *ec forsacho*, *gelöbo*. It is likewise absent from Low Dutch. Old West Saxon here diverges from Frankish in so far as it knows this *-n* in only one conjugation (the so-called second weak): *skawön* I look, *thionön* I serve, etc. It is completely foreign to Anglo-Saxon and Old Frisian. We may therefore presume that this *-n* too is a Frankish remnant in Old West Saxon.

Apart from the numerous proper names preserved to us in documents etc. and the often beyond recognition disfigured glosses of the Lex Salica, we have almost no remnants of the Salian dialect. However, Kern (*die Glossen in der Lex Salica*) has removed a considerable number of these disfigurements and restored the text, in some cases certain, in others highly probable, and shown that it is written in a language which is the direct ancestor of Middle and Modern Low Dutch. But this material thus reconstructed is of course not directly usable for grammar. In addition, we possess only the short renunciation formula appended to the Capitulary of Carloman of the year 743 and probably drawn up at the Council of Lestines, hence in Belgium. And here right at the beginning we encounter two characteristically Frankish words: *Ec forsacho*, I renounce. – *Ec* for I is still widespread today among Franks. In Trier and Luxembourg *eich*, in Cologne and Aachen *éch*, in Berg *ek*. While written Low Dutch has *ik*, one yet often enough hears in popular speech, especially in Flanders, *ek*. The Old Salian names *Segenandus*, *Segemundus*, *Segefredus* show uniformly *e* for *i*.

[1] *Kleine altsächsische und altniederfränkische Grammatik* by Moritz Heyne. Paderborn 1873.

[378] Frankish Period

In *forsacho*, ch stands for g between vowels: this occurs also elsewhere in the monuments (*rachineburgius*) and is even today a characteristic of all Frankish dialects from the Palatinate to the North Sea. We shall return to these two main characteristics of Frankish: e frequently for i, and ch between vowels for g, in connection with the individual dialects.

As a result of the above investigation, to which one may also compare what Grimm says about Old Frankish at the end of the first volume of the *History of the German Language*, we may lay down the proposition—which, by the way, is nowadays hardly contested any more—: that as early as the 6th and 7th centuries, Frankish was a distinct dialect, forming the transition between High German, i.e., primarily Alamannic, and Ingvaeonic, i.e., primarily Saxon and Frisian, and at that time still standing entirely at the Gothic-Low German stage of consonantal shift. But if this is admitted, then it is also recognised that the Franks were not a hotchpotch of different tribes thrown together by external circumstances, but a distinct main German tribe, the Iscaevones, who, although they at various times absorbed alien elements, also had the strength to assimilate them. And we may likewise regard it as proved that each of the two main branches of the Frankish tribe spoke a distinct dialect at an early date, that the dialect split into Salic and Ripuarian, and that many differentiating peculiarities of the old dialects still survive in the present-day vernacular.

Let us now pass on to these still-living dialects.

I. There is no longer any doubt that Salian lives on in the two Netherlands dialects, Flemish and Dutch, and indeed in its purest form in those areas that have been Frankish since the 6th century. For since the great storm floods of the 12th, 13th and 14th centuries destroyed almost the whole of Zeeland, created the Zuider Zee, the Dollart and the Jahde, and thereby broke the geographic [38] as well as the political cohesion among the Frisians, the remnants of the old Frisian liberty succumbed to the encroachments of the surrounding territorial lords, and with it, almost everywhere, the Frisian language as well. In the west it was hemmed in by Dutch, in the east and north by Saxon and Danish, or wholly displaced, in all cases leaving strong traces in the intruding language. The old Frisian Zeeland and Hol[378]land became in the 16th and 17th centuries the core and backbone of the Dutch struggle for independence, just as they had already been the seat of the chief trading cities of the country. Here, therefore, predominantly, New Dutch as a written language developed, and it absorbed Frisian elements, words and word-forms, which are to be clearly distinguished from the Frankish substratum. On the other hand, Saxon speech has advanced from the east into territory that was formerly Frisian and Frankish. Drawing the precise borders must be left to detailed research; purely Salian are only the Flemish-speaking parts of Belgium, North Brabant, Utrecht, as well as Gelderland and Overijssel, with the exception of the eastern, Saxon strips.

Between the French linguistic border along the Meuse and the Saxon border north of the Rhine, Salians and Ripuarians meet. We shall return to the dividing line, which here too can only be determined in detail, further below. Let us first deal with the grammatical peculiarities of Dutch.

Among the vowels, the first thing that strikes one is that in true Frankish fashion i is replaced by e: brengen bringen, kreb Krippe, hemel Himmel, geweten Gewissen, ben bin, stem Stimme. This is still far more frequently the case in Middle Dutch: gewes gewiß, es ist, selver Silber, blent blind, where New Dutch has gewis, is, zilver, blind. Likewise, I find near Ghent two places: Destelbergen and Desteldonck, which shows that even today Distel is called Destel there. Middle Dutch, which grew up on purely Frankish soil, corresponds here exactly to Ripuarian, while the written New Dutch, exposed to Frisian influence, does so to a lesser degree.

Furthermore, again in agreement with Ripuarian, o appears instead of u before m or n followed by a consonant, though not as consistently as in Middle Dutch and Ripuarian. Alongside konst, gonst, kund, New Dutch has kunst, gunst, kund; on the other hand, both agree in [39] mond Mund, hond Hund, jong jung, ons uns.

In contrast to Ripuarian, the long i (ij) has become ei in pronunciation, which does not seem yet to have been the case in Middle Dutch. But this ei is not pronounced like High German ei = ai, but as a real e + i, though not quite as thin as, for example, ej in Danish and Slavic. The diphthong written not as i j but as ei sounds only slightly different from this. Correspondingly, for High German au there appears ou, ou.

Umlaut has disappeared from inflection. In declension, singular and plural, and in conjugation, indicative and subjunctive, have the same root vowel. On the other hand, umlaut appears in word-formation in a double form: 1) in the change of a to e through i, common to all post-Gothic dialects, [379] 2) in a form peculiar to Dutch, which developed only later. Middle Dutch, like Ripuarian, still knows hus Haus, eran braun, ram geräumig, tan Zaun, pluralis huse, brune. New Dutch knows only the forms huis bruin ruim tuin (ui = High German eu), which are alien to Middle Dutch and Ripuarian. On the other hand, eu forces its way in for short o (High German u) already in Middle Dutch: jeughet alongside joghet, New Dutch jeugd Jugend; doghet Tugend, dor Thür, kor Wahl, alongside which the forms with eu occur, while New Dutch recognises only deugd, keur, deur. This entirely corresponds to the eu for stressed Latin o that developed in Northern French from the 12th century onwards. Kern draws attention to a third case: in New Dutch, ei is an umlaut of ê (ee). All these three forms of umlaut are unknown to Ripuarian as well as to the other dialects, and constitute a special characteristic of Dutch.

Aid, alt, old, oit, uld, ult change into oud, out. This change is already found in Middle Dutch, where, however, guldin, hulde, sculde still occur alongside goudin, houde, scoude (sollte), so that the time when it was introduced can be determined approximately. It is likewise peculiar to Dutch, at least in relation to all continental Germanic dialects; on the other hand, it also exists in the English dialect of Lancashire: gowd, howd, owd, for gold, hold, old.

As for the consonants, Dutch knows no pure g (the guttural [40] Italian, French or English g). This consonant is pronounced like a strongly aspirated gh, which in certain sound combinations is indistinguishable from the deep-guttural (Swiss, Modern Greek or Russian) ch. We have seen that this shift from g to ch was already known to Old Salian. It is also found in a part of Ripuarian and in those Saxon dialects that developed on formerly Frankish soil, e.g., in the Münsterland, where, as also in the Bergisches Land, j at the beginning of words, especially in loan-words, under certain conditions sounds like ch, and one can hear Choseph and even dar Chahr (Jahr). If M. Heyne had taken this into consideration, he would largely have been spared the difficulty of the frequent confusion and mutual alliteration of j, g, and ch in the Heliand.

At the beginning of words, Dutch retains wr in some places: wringen ringen, wreed grausam, hart, wreken rächen. A remnant of this remains also in Ripuarian.

From Frisian is taken the softening of the diminutive -ken into -tje, -je: mannetje Männchen, bier je Bienchen, halsje Hälschen, etc. However, k is also retained: vrouken Frauchen, hoedeken Hütchen. Flemish, at least in the vernacular, preserves the k better; the familiar little man [381] in Brussels is called manneken pis. Thus it is from Flemish that the French have borrowed their mannequin, and the English their mannikin. The plural of both endings is -s: vroukens, mannetjes. We shall encounter this -s again in Ripuarian.

Dutch has in common with Saxon and even Scandinavian dialects the expulsion of d between vowels, especially between two e’s: leder and leer, weder and weer, neder and neer, vader and vaer, moeder and moer Mutter.

Dutch declension shows a complete intermixture of strong and weak forms, so that, since plural umlaut is also absent, Dutch plural formations agree even with those of Ripuarian or Saxon only in the rarest cases, and here too an easily grasped characteristic of the language presents itself.

Common to Salian and Ripuarian, together with all Ingvaeonic dialects, is the omission of the nominative sign in er, der, wer: Dutch hij, de (article) and die (demonstrative pronoun), wie.

To go into the conjugation would lead too far. What has been said will suffice [41] to allow today’s Salian speech to be distinguished everywhere from the bordering dialects. Closer investigation of the Dutch vernaculars will assuredly bring to light much that is important.

II. Rhine Frankish. By this term I designate all the remaining Frankish dialects. The fact that I do not here contrast Salian, in the old way, with Ripuarian, has a very good reason.

Arnold has already pointed out that the Ripuarians, in the strictest sense, occupied a relatively narrow area, whose southern border is more or less marked by the two places Reifferscheid near Adenau and near Schleiden. This is correct in so far as it also demarcates linguistically the purely Ripuarian area from those territories occupied by genuine Ripuarians after, or simultaneously with, other German tribes. Since the name Lower Frankish has meanwhile acquired a different meaning that also includes Salian, the only term left to me for designating the group of closely related dialects extending from the Salian linguistic boundary to this line is Ripuarian—in the narrower sense.

1. Ripuarian. The dividing line of this group of dialects from Salian in no way coincides with the Dutch-German frontier. On the contrary, on the right bank of the Rhine, the greater part of the district of Rees still belongs to Salian, where in the area of Wesel Salian, Ripuarian, and Saxon meet. On the left bank, Cleves and Guelders are Salian up to approximately a line drawn from the Rhine, between Xanten and Wesel, southwards to the village Vluyn (west of Moers) and from there south-westwards towards Venlo. Exact determination of the boundary is only possible on the spot, [42] because, as a result of many years of Dutch administration, not only of Guelders but also of the county of Moers, many Ripuarian names have retained a Salian-Dutch form on the maps.

From the region of Venlo upwards, the greater part of the right bank of the Meuse appears to be Ripuarian, so that the political frontier here nowhere cuts through Salian, but consistently Ripuarian territory, and this extends to near Maastricht. Names in -heim (not -hem) and in the specifically Ripuarian -ich occur here in great numbers on Dutch territory, further south already those in -broich (Dutch -broek), e.g. Dallenbroich near Roermond; likewise in -rade (Bingelrade near Sittard, together with Amstenrade, Hobbeirade and 6—7 others); the little piece of German territory that fell to Belgium on the right of the Meuse is entirely Ripuarian (cf. Krüllenberg 9 kilometres from the Meuse with Kruisberg north of Venlo). Indeed, on the left of the Meuse, in Belgian so-called Limburg, I find Kessenich near Maaseyk, Stockheim and Reekheim on the Meuse, Gellik near Maastricht as proof that no purely Salian population lives here.

Towards Saxony the Ripuarian frontier runs from the area of Wesel southeastwards at an increasing distance from the Rhine, between Mülheim on the Ruhr and Werden on the Frankish —, and Essen on the Saxon side, through to the frontier of Berg and Mark, here still the boundary of the Rhine Province and Westphalia today. It leaves this only south of Olpe, where it goes eastward, separating the Siegerland as Frankish from the Saxon Sauerland. Further east, the Hessian dialect soon begins.

The southern frontier mentioned above, against the dialect designated by me as Middle Franconian, coincides approximately with the southern boundaries of the old pagi of Avalgau, Bonngau and Eiflia, and goes from there westwards to the Walloon, keeping rather somewhat to the south. The area thus circumscribed comprises the old great pagus of Ribuaria together with parts of the pagi adjoining it to the north and west.

As already stated, Ripuarian agrees with Dutch in many respects, yet in such a way that Middle Dutch stands closer to it than Modern Dutch. With the latter, Modern Dutch, the Ripuarian pronunciation of ei — e + i and ou for au agrees, the transition of i into e, which in Ripuarian, as in Middle Dutch, goes still much further than in Modern Dutch: the Middle Dutch gewes, es, blend, selver (silver) are still good Ripuarian today. Likewise, u before m or y with a following consonant changes consistently into o: jong, lomp, domm, konst. If this following consonant is a d or t, it changes in some dialects into g or k; e.g. honk hound, plural häng, where the softening to g is an after-effect of the dropped final vowel -e.

On the other hand, the umlaut relations of Ripuarian are sharply distinguished from the Dutch ones; on the whole they agree with the High German ones, and in individual exceptions (e.g. hanen for Hähne) with the Saxon ones.

In the initial position, wr has hardened to fr, preserved in fangen (wringen), to wring water from a cloth, etc., and fréd (Dutch wreed) with the meaning hardened, inured.

For er, der, wer stands hé, de, wé.

The declension stands midway between High German and Saxon. Plural formations in -s are frequent, but almost never agree with the Dutch ones; this -s becomes -r in the local High German, in correct recollection of the linguistic development.

The diminutive -ken, -chen changes after n into -schen, männschen; the plural has, as in Dutch, -s (männches). Both forms we will follow as far as Lorraine.

r before s, st, d, t, z is dropped; the preceding vowel remains short in some dialects, in others it is lengthened. Thus hart becomes hatt (Bergisch) haad (Cologne dialect). At the same time, through Upper German influence, st becomes scht: Durst, doascht (Bergisch) döscht (Cologne dialect).

Likewise, through High German influence, initial sl, sw, st, sp has become schl, etc.

Like Dutch, Ripuarian has no pure g. A part of the dialects lying on the Salian frontier, as well as the Bergisch dialect, likewise has aspirated gh for initial and medial g, but softer than the Dutch. The rest have j. In the final position, g is everywhere pronounced like ch, but not the hard Dutch ch, rather the softer Rhenish Franconian ch, which sounds like a hardened j. The essentially Low German character of Ripuarian is testified by expressions such as boven for oben.

The majority of the mute consonants stand everywhere on the first stage of the sound shift. Only in the case of t and of medial and final k and sometimes p has the High German shift taken place for the southern dialects; they have lôsze for löten (let, lassen), holz instead of holt, rieh instead of rik (reich), Ech instead of ek (ich), pief instead of pipe (Pfeife). But et, dat, wat and some others remain.

Indeed, it is precisely this, not even consistently carried out, penetration of the High German sound shift in three cases, on which the usual delimitation of Middle and Low Franconian is based. But by this means a group of dialects that are interconnected by certain phonetic relations, as shown, and that still recognise themselves in the popular consciousness as belonging together, is torn apart arbitrarily and according to a feature that is here entirely accidental.

Entirely accidental, I say. The other Central German dialects, the Hessian, Thuringian, Upper Saxon, etc., each stands on a stage of the High German shift that is on the whole definite. They may show somewhat less shift on the Low Saxon frontier, somewhat more on the Upper German frontier, but that at most gives rise to local differences. In contrast, Frankish shows no shift at all on the North Sea, Meuse and Lower Rhine, almost entirely Alemannic shift on the Alemannic frontier, and in between lie at least three intermediate stages. The sound shift has thus penetrated into the already independently developed Rhenish Franconian and has torn it into several pieces. The last trace of this shift by no means needs to disappear at the frontier of a previously existing, distinct group of dialects; it can die out in the middle of such a group, and indeed it does so. On the other hand, the genuinely dialect-forming influence of the sound shift, as will be shown, certainly ceases at the boundary of two dialectal groups that were already different earlier. And did not the schl, schw, etc., the final scht, likewise come to us from the High Germans, and far later? But this — at least the former — still goes deep into Westphalia.

The Ripuarian dialects formed a firm group long before a part of them learned to shift t, medial and final k and p. How far this change could penetrate within the group was and remains for the group purely accidental. The dialect of Neuss is, down to minutiae which a foreigner does not hear at all, identical with that of Krefeld and München-Gladbach. In spite of this, the one is supposed to be Middle Franconian, the other Low Franconian. The dialect of the Bergisch industrial country passes over in imperceptible gradations into that of the southwestern plain of the Rhine. And yet they are supposed to belong to two fundamentally different groups. For anyone who is at home in the region itself it is evident that here armchair scholarship forces the living folk dialects, which it knows little or not at all, into the Procrustean bed of a priori constructed characteristics.

And where does this purely external distinction lead? To this, that one lumps together the southern Ripuarian dialects into a so-called Middle Franconian with other dialects from which they, as we shall see, deviate far more widely than from the so-called Low Franconian ones. And that on the other hand one retains a narrow strip with which one does not know what to do, and one finds oneself finally compelled to declare one piece to be Saxon, a second to be Dutch, which flies directly in the face of the facts of these dialects.

Take, for example, the Bergisch dialect, which Braune summarily and confidently calls Saxon. It forms, as we saw, all three plural persons of the present indicative alike, but in the Frankish manner in the ancient form -nt. It regularly has ||[46]| o instead of u before m and y with a following consonant, which according to the same Braune is decidedly un-Saxon and specifically Low Franconian. It agrees in all the characteristics cited above with the other Ripuarian dialects. While it passes imperceptibly from village to village, from farmstead to farmstead into the dialect of the Rhine plain, it is divided from the Saxon dialect along the Westphalian frontier with razor-sharpness. Perhaps nowhere else in the whole of Germany is there a linguistic frontier drawn with such immediacy as here. And what a difference in the language! The entire vowel system is as if turned upside down; the sharp Low Franconian ei confronts the broadest ai without mediation, as does ou confront au; of the many diphthongs and vowel off-glides not a single one coincides; here seh as in the rest of Germany, there s-ch as in Holland; here wi hant, there wi hebbed; here the dual forms gei and enk (you and you) used as plurals, there only i, ji and è, jü; here the sparrow is called, in common Ripuarian, Masche, there in common Westphalian Lüning. Not to speak of other peculiarities specifically proper to the Bergisch dialect, which likewise suddenly vanish at this boundary line.

To the foreigner, the peculiarity of a dialect becomes most palpable when the person concerned does not speak the dialect, but the High German that is more comprehensible to him, which indeed with us Germans usually takes place under the strong influence of the dialect. But in that case the allegedly Saxon inhabitant of the Bergisch industrial district is for the non-native absolutely indistinguishable from the inhabitant of the Rhine plain, who is supposed to be Middle Franconian, except by the somewhat harder aspirated gh where the other speaks j. The Bergisch Heckinghauser (from Oberbarmen, on the left of the Wupper), however, and the Märkisch Langerfelder, who lives barely a kilometre further east, stand even in the local High German of everyday life further apart from one another than the Heckinghauser and the Koblenzer, let alone the Aachener or Bonner.

To the Rhenish Frank himself, the penetration of the shift of t and final k makes so little impression of a linguistic divide that, even in an area quite familiar to him, he will have to stop and think where the boundary between t and z, k and ch lies, and that when crossing this boundary the one is almost as much to his taste ||[47]| as the other. This is facilitated still further by the many High German words with shifted sz, z, ch and t which have pressed into the dialects. A striking example is provided by the old Bergisch judicial ordinance from the 14th century (Lacomblet, Archiv, I p.79ff.) Here are found zo, uiss (aus) zween, be-

zahlen; beside it in the same sentence rarely, dat nutteste (nutzeste); similarly Dache, redelich beside reicket (reicht); Upiaden, upheven, hulper (Helfer) beside verkouffen. In another paragraph p. 85 even alternates zo and tho zu. In short, the dialects of the mountains and the plain continually intermingle, without this bothering the writer in the least. As always, this last wave with which the High German consonant shift overflows Frankish territory is also the weakest and shallowest. It is certainly of interest to mark the line up to which it reaches. But this line cannot be a dialect boundary; it cannot tear apart an independent group of old and closely related dialects from one another, and provide the pretext by virtue of which one would assign the violently separated fragments, in contradiction to all linguistic facts, to more distant groups.

2. Middle Franconian. From the foregoing it goes without saying that I set the northern boundary of Middle Franconian considerably further south than is usually done.

From the fact that the left-Rhenish Middle Franconian tract of country seems to have been in Alemannic possession at the time of Clovis, Arnold takes occasion to examine the local place names for traces of Alemannic settlement, and arrives at the result that a pre-Frankish, Alemannic population can be shown as far as the line Cologne—Aachen; whereby, self-evidently, the traces, most numerous in the south, become ever rarer towards the north. The place names, he says, point “to a temporary advance of the Alemanni even beyond the region around Koblenz and Aachen, as well as to a longer possession of the Wetterau and the southern areas in Nassau. For the names with the genuinely Alemannic endings -ach, -brunn, -felden, -hofen, -ingen, -schwand, -Stetten, -wangen and -weiler, which occur nowhere in purely Frankish territory, are found scattered from Alsace across the entire Palatinate, Rhenish Hesse and Rhenish Prussia, only that they become ever rarer towards the north, and more and more give way to the preferentially Frankish names in -bach, -berg, -dorf, -born, -feld -hausen, -heim and -scheid.” (Deutsche Urzeit, p.[140].) 

|[48]| Let us first examine the allegedly Alemannic names of the Middle Franconian country. The endings -brunn, -Stetten, -felden, -wangen have nowhere occurred to me on Reymann’s map (which, let it be said once for all, I use here). The ending -schwand occurs once: Metzelschwander Hof near Winnweiler, and then also Schwanden north of Landstuhl. Thus both times in the Upper Franconian Palatinate, which does not concern us here. Along the Rhine we have Kreuznach, Bacharach, Hirzenach near Sankt Goar, Rübenach near Koblenz, (Ribiniacus of Spruner-Menke’s gau map) Andernach (Antunnacum of the Romans) beside Wassenach. Since on the entire left bank of the Rhine the Romanised Celtic ending -acum occurs generally in Roman times — Tolbiacum Zülpich, Juliacum Jülich, Tiberiacum Ziewerich near Bergheim, Mederiacum — in most of these cases only the choice of the form -ach instead of -ich could at most indicate Alemannic influence. Only the one Hirzenach (= Hirschenbach) is unconditionally German, and this, according to the gau map, was earlier called Hirzenowe, Hirschau, not Hirschenbach. But how then explain Wallach, which lies between Büderich and Rheinberg, hard on the Salian boundary? That is surely not Alemannic.

In the Mosel area a few -ach also occur: Irmenach east of Bernkastel, Waltrach, Crettenach near Trier, Mettlach on the Saar. In Luxembourg Echternach, Medernach, Kanach; in Lorraine only on the right of the Mosel: Montenach, Rodelach, Brettnach. Even if we were to concede that all these names point to Alemannic settlement, then only to a very thinly sown one, which moreover does not go beyond the southernmost part of the Middle Franconian country.

There remain -weiler, -hofen and -ingen, which require more thorough investigation.

The ending -weiler is first of all not without further ado Alemannic, but the provincial-Latin villarium, villare and is found at most quite exceptionally outside the ancient boundaries of the Roman Empire. It was not the German rendering of villare into weiler that was the privilege of the Alemanni, but only the predilection with which they applied this ending en masse for new settlements as well. In so far as Roman villae occurred, the Franks too were compelled to render the ending as wilare, later ||[49]| weiler, or to drop it entirely. Probably they did now the one, now the other, just as they will certainly also have given names in -weiler to new settlements here and there, only far more rarely than the Alemanni.

North of Eschweiler near Aachen and Ahrweiler Arnold can find no more important places in -weiler. But the present importance of the places is beside the point; the fact is that on the left bank of the Rhine the -weiler extend northwards nearly as far as the Salian boundary (Garzweiler and Holzweiler are not five miles from the nearest Dutch-speaking locality of Gelderland) and north of the line Eschweiler—Ahrweiler there are at least twenty of them. They are most frequent, understandably, in the vicinity of the old Roman road from Maastricht via Jülich to Cologne, two of them, Walwiller and Nyswiller, even on Dutch territory — are these also Alemannic settlements?

Further south they hardly occur at all in the Eifel; the Malmedy section (No. 159 Reymann) has not a single instance. In Luxembourg they are also rare; likewise on the lower Mosel and up to the crest of the Hunsrück. On the other hand they appear frequently on the upper Mosel on both sides of the river, becoming ever denser towards the east, and east of Saarlouis they become more and more the dominant ending. Here, however, Upper Franconian speech already begins, and here no one disputes that the Alemanni had occupied the land before the Franks.

For the Middle Franconian and Ripuarian territory, therefore, the Weiler prove no more Alemannic settlement than the many -villers in France.

Let us pass on to -hofen. This ending is all the less exclusively Alemannic. It occurs in the entire Frankish territory, including what is now Westphalia, later occupied by the Saxons. On the right bank of the Rhine only a couple of examples: Wehofen near Ruhrort, Meilinghofen and Eppinghofen near Duisburg, Benninghofen near Mettmann, another Eppinghofen near Dinslaken, in Westphalia Kellinghofen near Dorsten, Westhofen near Castrop, Wellinghofen, Wichlinghofen, Niederhofen, two Benninghofen, Berghofen, Westhofen, Wandhofen, all on the Hellweg, etc. Ereshofen on the Agger, Mortis villa, reaches back into heathen times, and the very designation of the war god as Ero shows that no Alemanni are conceivable here; they called themselves Tiuwâri, thus called the god not Ero but Tiu, later shifted Ziu.

|[50]| On the left bank of the Rhine matters are even worse with the Alemannic descent of -hofen. There is again an Eppinghofen south-east of Xanten, thus perhaps already Salian, and from there southwards the whole Ripuarian territory teems with -hofen, alongside -hof for single farms. If we then go on to Salian land, it gets still worse. The Meuse is accompanied on both sides, from the French language boundary onwards, by -hofen; for brevity’s sake let us go straight to the western bank. There we find in Holland and Belgium at least seven Ophoven, in Holland Kinckhoven, etc.; for Belgium let us first take the Löwen section (No. 139 Reymann). Here there are Ruykhoven, Schalkhoven, Bommershoven, Wintershoven, Mettecoven, Heishoven, Engelmannshoven near Tongeren; Zonhoven, Reekhoven, Konings-Hoven near Hasselt, further west Bogenhoven, Schuerhoven, Nieuwenhoven, Gippershoven, Baulershoven near St. Truiden; furthest west Gussenhoven and Droenhoven east and north-east of Tienen (Thienen). The Turnhout section (No. 120) has at least 33 -Hoven, most of them on Belgian territory. Further south-west the -hove (the dative -n is regularly suppressed here) run all along the French language boundary: from Heerlinkhove and Nieuwenhove near Ninove, which is itself a Romanised -hove, — I omit the intermediate links, some 10 counted — as far as Ghyverinckhove and Pollinchove near Diksmuide and Volckerinckhove near Saint Omer in French Flanders. Nieuwenhove occurs three times, which proves that the ending is still alive among the people. Alongside, very numerous single farms in -hof. After this let the allegedly exclusively Alemannic character of -hofen be judged.

Finally to -ingen. The designation of common descent by -ing, -ung, is common to all Germanic peoples. Since settlement took place by clans, this ending plays an important role everywhere in place names. Sometimes it is linked in the genitive plural with a local ending: Wolvarad-inga-husun near Minden, Snotingaham (Nottingham) in England. Sometimes the plural stands alone for the place-name: Flissingha (Vlissingen), Phladirtinga (Vlaardingen), Crastlingi in Dutch Friesland; Grupilinga, Britlinga, Otlinga in Old Saxony. These names are nowadays mostly reduced to the dative and end in -ingen, rarely -ing. Most peoples know and use both modes of use; the Alemanni seem, it appears, predominantly the latter, at least at present.*) Since, however, this also occurs among Franks, Saxons and Frisians, it is very bold to conclude immediately from the occurrence of place names in -ingen an Alemannic settlement.

The names just cited prove that names in -ingas (nominative plural) and -ingum, -ingon (dative plural) were nothing out of the ordinary among both ||[51]| Frisians and Saxons from the Scheldt to the Elbe. Even today the -ingen are no rarity in all of Lower Saxony. In Westphalia, on both sides of the Ruhr, south of the line Unna—Soest, there are alone at least twelve -ingen, alongside -ingsen and -inghausen. And as far as Frankish territory extends, so far do we also find names in -ingen.

On the right bank of the Rhine we find first in Holland Wageningen on the Rhine and Genderingen on the IJssel (whereby we exclude all possibly Frisian names), in the Bergisches Land Huckingen, Ratingen, Ehingen (right behind it on Saxon territory Hattingen, Sodingen, Ummingen), Heisingen near Werden (which Grimm derives from the Silva Caesia of Tacitus, and which would thus be age-old), Solingen, Husingen, Leichlingen (on the gau map Leigelingon, thus about 1,000 years old), Quettingen, and on the Sieg Bödingen and Röcklingen, not counting two names in -ing. Hönningen near Rheinbrohl and Ellingen in the Wied region establish the connection with the district between the Rhine, Lahn and Dill, which, conservatively counted, yields 12 -ingen. To go further south is pointless, since here the land begins
that undisputedly went through a period of Alemannic settlement.

*) Rümmingen near Lörrach was earlier called (764) Romaninchova, so that the Swabian -ingen are sometimes also of more recent origin. (Mone, Urzeit des badischen Landes, I, p.213) The Swiss -kon and -kofen are almost all contracted from -inghofen: Zollinchovun, Zollikhofen, Smarinchova Schmerikon etc. Cf. Friedrich Beust, historischer Atlas des Kantons Zürich, where they are found by the dozen on Map 3 representing the Alemannic period.

390

The Frankish Period

On the left bank of the Rhine we have Millingen in Holland above Nijmegen, Lüttingen below Xanten, again Millingen below Rheinberg, then Kippingen, Rödingen, Höningen, Worringen, Fühlingen, all further north than Cologne, Wesselingen and Köttingen near Brühl. From here the names in -ingen pursue two directions. In the high Eifel they are rare; we find near Malmédy on the linguistic frontier: Büllingen, Hünningen, Mürringen, Iveldingen, Eibertingen as a transition to the very numerous -ingen in Luxembourg and on the Prussian and Lotharingian upper Moselle. Another line of connection runs along the Rhine and its lateral valleys (in the Ahr region 7–8), finally along the Moselle valley likewise into the region above Trier, where the -ingen predominate, but are first separated from the great mass of Alemannish-Swabian -ingen by the -weiler, and then the -heim. If therefore we ||[52]] “take all circumstances into consideration in their context,” as Arnold demands, we shall arrive at the conclusion that the -ingen of the upper German Moselle region are Frankish and not Alemannish.

How little we need Alemannish help here becomes completely clear as soon as we follow the -ingen from the Franco-Ripuarian linguistic frontier near Aachen into Salian territory. At Maaseyk west of the Meuse lies Geystingen, further west near Bree Gerdingen. Then we find, if we again take up sheet No. 139, Löwen, the following: Mopertingen, Vlytingen, Rixingen, Aerdelingen, Grimmersingen, Gravelingen, Ordange (for Ordingen), Bevingen, Hatingen, Buvingen, Hundelingen, Bovelingen, Curange, Raepertingen, Boswinningen, Wimmertingen and others, in the region of Tongern, St. Truyen and Hasselt. The westernmost, not far from Löwen, are Willebringen, Redingen, Grinningen. Here the chain appears to break off. But if we turn to the region that now speaks French but was contested between the two languages in the 6th–9th centuries, we find from the Meuse onwards a whole belt of Frenchified -ange, which form also corresponds to -ingen in Lothringen and Luxembourg: going from east to west: Ballenge, Roclenge, Ortrange, Lantremange, Roclange, Libertange, Noderange, Herdange, Oderinge, Odange, Gobertang, Wahenges; somewhat further west, Louvrenge near Wavre and Revelinge near Waterloo establish the link with Huysinghen and Buisinghen, the first posts of a group of over 20 -inghen which are spread along the linguistic frontier south-west of Brussels from Hai to Grammont. And finally, in French Flanders: Gravelingen, Wulverdinghe (thus entirely the Old Saxon Wolvaradinges-Msun), Leubringhen, Leulinghen, Bonninghen, Peuplingue, Hardinghen, Hermelinghen, near St. Omer and as far as beyond Boulogne Herbinghen, Hocquinghen, Velinghen, Lottinghen, Ardinghen, all sharply distinct from the names in -inghem (= -ingheim) which are still more numerous in the same region.

The three endings, then, which Arnold holds to be specifically Alemannish, have proved to be just as much Frankish, and the attempt to prove an Alemannish ||[53]] settlement on Middle Frankish territory, antecedent to the Frankish one, from these names must be regarded as a failure. The possibility of a not very strong Alemannish element in the south-easternmost part of this area may always be admitted, however.

From the Alemanni Arnold leads us to the Chatti. The latter, to the exclusion of the Ripuarians proper, are said to have occupied the territory south of the Ripuarian Gau—that is, the same which we call Middle and Upper Frankish—after and alongside the Alemanni. This argument is also based on the Hessian place-names found in this region alongside the Alemannish ones: “The correspondence of place-names on this side and the other side of the Rhine as far as the Alemannic frontier is so remarkable and striking that it would be a veritable miracle if it were accidental, whereas it appears quite natural as soon as we assume that the immigrants likewise gave their native place-names to their new settlements, as happens every day in America.”

There is little to be said against this proposition. All the more against the conclusion that the Ripuarians proper had nothing to do with the settlement of the whole Middle and Upper Frankish land, that we find here only Alemanni and Chatti. Most of the Chatti who moved westward from their homeland seem to have attached themselves at all times (as already the Batavi, Caninefates and Chattuarii) to the Istvaeones; where else were they to go? In the first two centuries of our era, the Chatti were linked with the other Herminones only in the rear through the Thuringians; on one side they had Ingvaeonic Cherusci, on the other Istvaeones, and in front of them the Romans. The Herminonian tribes which later appear united as Alemanni came from inner Germania, were separated from the Chatti for centuries by Thuringians and other peoples, and had become much more alien to them than the Istvaeonian Franks, with whom they were bound by many centuries of comradeship in arms. The participation of the Chatti in the occupation of the tract of land in question is thus not doubted, but the exclusion of the Ripuarians from it is. This ||[54]| exclusion can only be regarded as proved if no specifically Ripuarian names occur here. The reverse is the case.

Among the endings given by Arnold as specifically Frankish, -hausen is common to Franks, Saxons, Hessians and Thuringians; -heim is called -ham in Salian, -bach is Salian and Lower Ripuarian -beek; of the others only -scheid is really characteristic. It is specifically Ripuarian, as are -ich, -rath or -rade and -siepen. Common to both Frankish dialects are further -loo (-loh), -donk and -bruch or -broich (Salian broek).

Scheid occurs only in the mountains and as a rule from settlements on the watershed. The Franks have left this ending throughout the Westphalian Sauerland up to the Hessian frontier, where it appears, now only as mountain names, as far as east of Korbach. On the Ruhr the Old Frankish -scheid is confronted by the Saxonised ending -schede: Melschede, Selschede, Meschede, close by Langscheid, Ramscheid, Bremscheid. Frequent in the Bergisches Land, it is found as far as the Westerwald, but no farther south on the right bank of the Rhine. On the left bank of the Rhine, on the other hand, the -scheid understandably do not begin until the Eifel*); in Luxembourg there are at least 21 of them, they are frequent in the Hochwald and Hunsrück. But as south of the Lahn, so here too the form -schied appears alongside them on the east and south side of the Hunsrück and Soonwald, which seems a Hessian adaptation. The two forms side by side extend southwards across the Nahe into the Vosges, where we find: Bisterscheid west of the Donnersberg, Langenscheid near Kaiserslautern, a plateau Breitscheid south of Hochspeier, Haspelscheid near Bitsch, the Scheidwald north of Lützelstein, finally as the southernmost post Walscheid on the northern slope of the Donon, still further south than the village of Hessen near Saarburg, the furthest Chatti post according to Arnold.

||[55]| Specifically Ripuarian further is -ich, from the same root as Gothic -ahva water, like -ach; both also Germanise the Belgo-Roman -acum, as is proved by Tiberiacum, on the Gau map Civiraha, today Ziewerich. On the right bank of the Rhine it is not very frequent; Meiderich and Lirich near Ruhrort are the northernmost, from there they run along the Rhine as far as Biebrich. The plain on the left bank of the Rhine, from Büderich opposite Wesel, is full of them, through the Eifel they go as far as the Hochwald and Hunsrück, but disappear in the Soonwald and the Nahe region even before -scheid and -roth cease. In the western part of our area, on the other hand, they continue up to the linguistic frontier and beyond. We shall pass over the Trier region, which has a great many; in Dutch Luxembourg I count twelve, further beyond in Belgian territory Türnich and Merzig (Messancy—the spelling -ig changes nothing, etymology and pronunciation are the same), in Lothringen Soetrich, Sentzich, Marspich, Daspich west of the Moselle; east of it Kuntzich, Penserich, Cemplich, Oestrich, twice Kerprich, Hibrich, Hilsprich.

The ending -rade, -rad, on the left bank of the Rhine -rath, likewise extends far beyond the boundaries of its old Ripuarian homeland. It fills the entire Eifel and the middle and lower Moselle valley as well as its lateral valleys. In the same region where -scheid mixes with -schied, there occur on both banks of the Rhine -rod, -roth alongside -rad and -rath, likewise of Hessian origin. Only that on the right bank, in the Westerwald, the -rod go further north. In the Hochwald, the northern slope regularly has -rath, the southern slope -roth.

Least advanced is -siepen, shifted -seifen. The word denotes a small brook valley with a steep gradient, and is still in general use for it. On the left bank of the Rhine it does not extend far beyond the old Ripuarian frontier; on the right it is found in the Westerwald on the Nister and still near Langenschwalbach (Langenseifen).

Going into the other endings would lead too far. ||[56]| In any case, however, we may declare the countless -heim that accompany the Rhine from Bingen upwards far into the Alemannish area, and which are found everywhere that Franks have settled, to be not Chatti but Ripuarian. Their homeland is not in Hesse, where they occur seldom and seem to have penetrated later, but in the Sauerland and the Rhine plain around Cologne, where they occur alongside the other, specifically Ripuarian names in almost equal numbers.

The result of this investigation is therefore that the Ripuarians, far from being pinned down to the Westerwald and Eifel by the stream of Hessian immigration, on the contrary themselves flooded the entire Middle Frankish territory. And indeed more strongly in the south-western direction, towards the upper Moselle region, than towards the south-east, towards the Taunus and the Nahe region. This will also be confirmed by the language. The south-western dialects as far as Luxembourg and West Lothringen are far closer to Ripuarian than the eastern ones, especially those on the right bank of the Rhine. The former may be regarded as a prolongation of Ripuarian with a stronger High German sound shift.

The characteristic feature of the Middle Frankish dialects is, first of all, the penetration of the High German sound shift. Not the mere shifting of some tenues to aspirates, which extends to relatively few words and does not affect the character of the dialect, but the incipient shifting of the mediae, which brings about the distinctive Middle and Upper German confusion of b and p, g and k, d and t. Only where it becomes impossible to distinguish sharply between b and p, d and t, g and k in the initial sound—that which the French especially understand by accent allemand—only there does the Low German become conscious of the great rift which the second sound shift has torn through the German language. And this rift runs between the Sieg and the Lahn, between the Ahr ]|[57]| and the Moselle. Accordingly, Middle Frankish has an initial g which is absent in the more northerly dialects, though in medial and final position it still pronounces soft ch for g. Furthermore, the ei and ou of the northern dialects pass over into ai and au.

*) (Note.) In the plain I find only Waterscheid east of Hasselt in Belgian Limburg, where we already observed strong Ripuarian admixture above.

Frankish Period

Some eight Frankish peculiarities: in all Salian and Ripuarian dialects *Bach*, unshifted *Beek*, is feminine. This also holds for at least the greater western part of Middle Frankish. Like the countless other brooks of the same name in the Netherlands and on the Lower Rhine, the Luxembourgish Glabach (Gladbach, Dutch Glabeek) is also feminine. On the other hand, girls’ names are treated as neuters: one says not only *das Mädchen*, *das Mariechen*, *das Lisbethchen*, but also *das Marie*, *das Lisbeth*, from Barmen to beyond Trier. — Near Forbach in Lorraine, the map—originally taken by Frenchmen—shows a “Karninschesberg” (Kaninchenberg). Thus the same diminutive *-sehen*, plural *-sches*, which we found above as Ripuarian.

With the watershed between Moselle and Nahe and, on the right bank of the Rhine, the hill country south of the Lahn, a new group of dialects begins:

3. Upper Frankish. Here we find ourselves on a tract of land that was indisputably at first Alemannic conquest territory (apart from the earlier occupation by the Vangiones, etc., of whose tribal affinity and language we know nothing) and where a stronger Chatti admixture may also readily be conceded. But here too the place names, as we need not repeat, point to the presence of not insignificant Ripuarian elements, especially in the Rhine plain. Still more so the language itself. Let us take the southernmost determinable dialect that at the same time possesses a literature, the Palatine. Here we again encounter the general Frankish impossibility of pronouncing in- and final *g* otherwise than as soft *ch.*) It is said there: *Vöchel, Flechel, ge-eche* (gelegen) *gsacht* gesagt, *licht* liegt, etc. Likewise the general Frankish *w* instead of *b* in medial position: *Büwe* Buben, *gläwe* glauben (aber *i gläb*), *bleiwe, selwer* selbst, *halwe* halbe. The shift is far from being as complete as it appears; indeed, especially in foreign words, a back-shift occurs, that is, the mute consonant of the initial position is shifted not one stage forward but backward, *t* becomes *d*, *o* becomes *b*, as will be shown; *d* and *p* in initial position remain on the Low German stage: *dün* than, *dag, danze, diir, dodt*; however, before *r: trinke, trage, paffidli, peife, palz* Pfalz, *parre* Pfarrer. Since *d* and *p* stand for High German *t* and *pf*, initial *ź* in foreign words is also back-shifted to *d*, and initial *po* back-shifted to *b: derke* Türke, *dafel* Tafel, *babeer* Papier, *borzlan* Porzellan, *bulwer* Pulver. Then the Palatine dialect—in this agreeing only with Danish—tolerates no tenues between vowels: *ebbes* etwas, *labbe* Lappen, *schlubbe* schlupfen, *schobbe* Schoppen, *Peder* Peter, *dridde* dritte, *rodhe* rathen. Only *k* is an exception: *brocke, backe*. But in foreign words *g: musigande* Musikanten. This is likewise a remnant of the Low German sound stage, which has extended itself further by means of back-shift; it was only because *dridde*, *hadde* remained unshifted that Peter could become *Peder*, and thus the corresponding High German *t* received similarly impartial treatment. Similarly, in *halde* halten, *aide* alte, etc., *d* remains on the Low German stage.

In spite of the decidedly High German overall impression—for Low Germans—the Palatine dialect is therefore far from having adopted the High German sound shift even to the extent that our written language preserves it. On the contrary, Palatine protests by means of its back-shift against the High German stage that has intruded from without and proves itself to this day a foreign element in the dialect.

This is the place to go into a generally misunderstood phenomenon: the confusion of *d* and *t*, *b* and *p*, even *g* and *k*, among those Germans in whose dialect the media have undergone the High German shift. This confusion does not occur as long as everyone speaks his own dialect. On the contrary. We have just seen that the Palatine, for example, distinguishes them very precisely, so much so that he even back-shifts foreign words in order to adapt them to the requirements of his dialect. Foreign initial *t* becomes *d* for him only because written German *t* corresponds to his *d*, foreign *p* becomes *b* because written German *pf* corresponds to his *p*. Likewise, the mute consonants are not jumbled together in other Upper German dialects as long as the dialect is spoken. Each of them has its own precisely applied shift law. Matters change as soon as the written language or a foreign language is spoken. The attempt to apply the current dialectal shift law to it—and this attempt is made involuntarily—collides with the attempt to speak the new language correctly. In the process, the written *b* and *p*, *d* and *t* lose all fixed meaning, and thus it could happen that Börne, for example, complains in his *Paris Letters* that the French cannot distinguish *b* and *p* because they stubbornly believe that his name, which he pronounced *Perne*, began with a *P*.

But back to the Palatine dialect. The proof that the High German shift has been forced upon it, so to speak, from without and has remained a foreign element to this day, and that it has not even reached the sound stage of the written language (far surpassing the latter, Alemannic and Bavarian on the whole preserve this or that Old High German stage)—this proof alone suffices to establish the predominantly Frankish character of Palatine. For even in Hesse, situated much further north, the shift is on the whole carried further, and thereby the allegedly predominantly Hessian character of Palatine is reduced to a more modest measure. In order to offer such resistance to the High German shift, hard on the Alemannic border, among backward Alemanni, alongside the Hessians—themselves essentially High German—at least as numerous Ripuarians must have been on the spot. And their presence is further proved—apart from the place names—by two generally Frankish peculiarities: the preservation of Frankish *w* instead of *b* in medial position, and the pronunciation of *g* as *ch* in medial and final position. To this are added a host of individual cases of agreement. With the Palatine *Gundach* (guten Tag), one gets as far as Dunkirk and Amsterdam. Just as in the Palatinate a *gewisser Mann* is called *ein sichrer Mann*, so in the whole Netherlands a *een zekeren man*. *Handsching* for Handschuh agrees with Ripuarian *Händschen*. Even *g* for *j* in *Ghannisnacht* (Johannisnacht) is Ripuarian, and extends, as we saw, to the Münsterland. And the word *baten* (bessern, nützen from *bat*, besser), common to all Franks including the Netherlanders, is current in the Palatinate: *’s badd alles nix*, es hilft alles nichts—where even the *t* is not shifted High German to *tz*, but is softened Palatine-style between vowels to *d*.

*) All citations are from “Fröhlich Palz, Gott erhalts! Gedichte in Pfälzer Mundart” by K. G. Nadler. Frankfurt am Main 1851.

Friedrich Engels