One of the most interesting, but sometimes slightly underestimated topics of research as a whole into the Late Antiquity of the ‘barbaric’ part of Europe is the development of longhouses and settlements. This paper is an attempt to combine the results of long-term research on construction and settlements from the Iron Age (with a main focus on the Roman Iron Age and Migration Period) in the western part of Central Europe and Scandinavia with the results of relevant research in Poland. This is no easy task. Despite undeniable research progress in recent decades, settlement archaeology in Poland is still in the early stage of searching for patterns of recognition and reconstruction of longhouses that can contribute to the determination of individual house types. The aim of this paper is to convince the Polish research community that it is necessary to change its perspective on the subject of Iron Age house building and especially on the spatial organisation of settlements. Too often, one can observe an avoidance of careful and accurate analysis of archaeological objects in relation to the reconstruction of house plans – partly out of fear of misinterpretation, partly due to inability, partly because of habit and use of well-worn research paths, but often also out of a lack of reflection on the regularities and laws of statics and carpentry methods. In this way (unnecessarily), a gap was created between two (artificially created) zones of barbaric Europe that lacks one of the basic features of working on archaeological material within the so-called Germania magna: comparability. For a long time, the pit house was regarded as the main residential building in Late Antiquity in the area of Poland. Additionally, post houses were and are being reconstructed that could never have existed in this way. As a result of efforts to adapt the shape of the house to his own needs and economic requirements, a man living in Central and Northern Europe had already created a universal building in the Neolithic (Fig. 2) that we call a longhouse. However, this building is not a homogeneous creation. In different periods of time, in regionally determined varieties, it occurs in different forms. On the basis of certain design features, arrangements of roof-bearing structures and other elements, these varieties are recognised as house types. Similarly to the classification of artefacts and analysis of the distribution of different types, variants and varieties, the analysis of house types also helps us to determine the peculiarities of individual societies and groups, to track their development and to recognise zones of common tradition and contact networks. At this point, I would venture to say that construction traditions even more closely reflect the characteristics of individual societies than, for example, brooches whose forms have undergone rapid fashion changes and influences from various milieus. For large areas in western Central Europe and Scandinavia, we can determine house types that can be grouped into overarching categories, defining building tradition zones (Hauslandschaften). In the relevant works, such regions east of the Oder have not yet found their place. It is high time to change that. I decided to review in the first part of the paper the most important issues related to Iron Age house building, given the fact that this paper cannot cover and discuss all aspects of the issue. Construction details, forms and basic types of longhouses in northern Central Europe are discussed, followed by the layout of farmsteads and settlements. The second part of the article attempts to relate the results of settlement archaeology in western Central Europe and Scandinavia to research results in Poland, often based on a reinterpretation of published features. When discussing the main features – the description of the post hole, the appearance and foundation of the post itself, the walls, doorways, roofs and house types, as well as the layout of farmsteads and settlements – I always had in mind and attempted to refer to the situation in Poland. It is a trivial statement that the most important feature in settlement research is the post hole. We owe the first detailed description of the archaeological feature which we call a post hole to A. Kiekebusch (1870–1935), an employee and later a department head of the Märkisches Museum in Berlin. He had contact with C. Schuchhardt (1859–1943), one of the founders of the Römisch-Germanische Kommission in Frankfurt am Main. From 1899, he, in turn, conducted excavations in the Roman legionnaire camp of the Augustus period in Haltern on the northern edge of the Ruhr region, during which, for the first time on a large scale, attention was paid to the remains of ancient post foundations. Thus, research in Haltern can be regarded as the beginning of modern settlement archaeology. During research on the early Iron Age stronghold Römerschanze in Potsdam, Schuchardt transferred the discovery of the research value of the post hole to ‘barbarian’ archaeology. The aforementioned A. Kiekebusch participated in research on Römerschanze; C. Schuchardt’s innovative research methods made a huge impression on him. In the publication of results of his own excavation of a Bronze Age settlement in Berlin-Buch, he described the appearance and properties of the post hole on eleven (!) pages (Fig. 4). The turn of the 19th/20th cent. is also a breakthrough in settlement archaeology in the Scandinavian countries. Here, however, the road was slightly different than on the continent, in a figurative sense from the general to the detail. Geographical conditions and construction methods, sometimes quite different from the way houses were erected in Central Europe, were conducive to the discovery of real Iron Age ruins of three-aisled houses and in this way it was known almost from the very beginning of settlement research that the houses were elongated and based on the structure of regularly placed roof-bearing posts. For example, in 1924, plans were published of the remains of burnt down houses in the Late Pre-Roman Iron Age settlement at Kraghede in northern Jutland that was discovered in 1906 (Fig. 5). The posts of these houses have survived partly as charred wood, which greatly facilitated the interpretation of discovered traces. The 1920s and 30s witnessed a real leap in settlement archaeology, which was also observed on the continent, e.g. in the Netherlands. A.E. van Giffen (1888–1973) conducted excavations in 1923–1934 in the area of the warf/Wurt/wierde/terp at Ezinge in the Dutch part of Friesland – a Late Pre-Roman and Roman Iron Age settlement. These names, mentioned in Dutch, Frisian and North German dialects, refer to an artificial hill in the North Sea shore region, created to protect house sites against high tide and floods. Moisture in the earth was conducive to the preservation of organic materials, and because of this van Giffen also found ‘real’ ruins of houses (Fig. 6). Large-scale excavations of this type in Germany were conducted in 1954–1963 at the Feddersen Wierde site. The results of this research were just as spectacular as in the case of the settlement at Ezinge (Fig. 46, 47). Large-scale research began in various countries in the 1960s as part of extensive research projects. In Denmark, the nationwide ‘Settlement and Landscape’ project resulted, among others, in the uncovering of a huge area with several settlements/farm clusters from the Pre-Roman Iron Age at Grøntoft, Jutland (Fig. 1). The completely surveyed, enclosed settlement from the Pre-Roman Iron Age at Hodde, Jutland must be mentioned in this context, too. At Vorbasse in Jutland, a huge area from the Late Roman Iron Age and Migration Period settlement was uncovered. After pioneering research at Feddersen Wierde in the 1970s, as part of the ‘North Sea Programme’ project of the Deutsche Forschungsgemeinschaft (German Research Community), research began at the 1st to 6th cent. CE settlement site at Flögeln in the German part of the southern coast of the North Sea. The results became fundamental not only for this region of Germany. As part of the competitive project ‘Research on Iron Age settlements’ of the Academy of Sciences in East Berlin, large-scale excavations were conducted in settlements of the Roman Iron Age and Migration Period settlements at Tornow in Lower Lusatia and at Herzsprung in the Uckermark. Already at the turn of the 1950s/60s, the famous Early and Late Roman Iron Age settlement at Wijster in the northern Netherlands was excavated, but the area studied was not comparable in size to the areas of the above-mentioned sites. In 1974, excavations began at Oss in the southern part of the country, starting in 1979 within the so-called Maaskant-Project of the University of Leiden, which led to the unveiling of an extremely large area, consisting of many, slightly dispersed excavations at so-called native settlements from the Pre-Roman Iron Age and the time when this region was part of the Roman Empire. North of the Rhine and Waal, in the northern Netherlands, the Peelo site is situated. Here, in the 1970s and 1980s, extensive excavations at several neighbouring settlement sites were carried out as part of the ‘Peelo project’ of the Biologisch-Archaeologisch Instituut of the University of Groningen. Similar large excavations were conducted in the 1980s at Colmschate in the eastern Netherlands by the Rijksdienst voor Oudheidkundig Bodemonderzoek, Archeologische Werkgemeenschap Nederland and Archeologie Deventer. The settlement traces date back to the Bronze Age up to medieval times. In the meantime, many new and important large-scale settlement excavations took place that cannot all be mentioned here. In the following chapters, I discuss the most important basic features of longhouses, beginning with the post hole and the post itself. Along with the growing sensitivity of archaeologists towards this issue and thanks to the good condition of surviving posts, there are more and more examples of houses where planks were used as roof-bearing poles. Excellent examples are the Late Pre-Roman Iron Age house at Jerup on Vendsyssel-Thy and two Late Roman Iron Age houses at Ragow and Klein Köris, both south of Berlin (Fig. 8). In some cases, there is evidence that the post was secured in the ground, such as a plank basement at the settlement of Klein Köris, anchoring at Feddersen Wierde or stones used as stabilisation like at Herzsprung (Fig. 7). In eastern Brandenburg, we have seen partial or complete post-hole fillings of burnt or unburnt clay, especially in the case of granaries. Depending on the function of the post, the sizes of the post holes can differ. The deepest post holes often belong to roof-bearing and doorway posts. It is interesting that this applies not only to three-aisle, but also to two-aisled houses (Fig. 10). This fact can be useful in the case of incomplete house plans. The basic typological division of longhouses refers to the general roof-bearing construction (three-aisled, two-aisled, one-aisled and so-called four-aisled houses). Three-aisled houses were not invented in the Iron Age; they appeared as early the Early Bronze Age (Fig. 11) within a large zone including northwestern France and Belgium, the Netherlands, Denmark and Sweden. Although closely related to the idea of keeping livestock in the same building where people lived, well-dated three-aisled houses with a stall do not date to earlier than around 1400 BCE. During the Pre-Roman and Roman Iron Age, the area of occurrence of these houses contracted slightly; they were erected in a wide zone south of the North Sea, in the Netherlands and northern Germany, Jutland, on the Danish islands and in southern areas of Norway and Sweden. Due to intensive settlement research carried out since the 1990s, we know that – at least in the Roman Iron Age – all of Mecklenburg, Western Pomerania, most of Brandenburg and some regions at the Middle Elbe belonged to this zone of three-aisled houses. The layout of two-aisled houses differs slightly due to construction based on only one row of roof-bearing posts. The arrangement and number of posts are often not as regular as in the case of three-aisled houses, which can create problems when interpreting house plans. Two-aisled longhouses, known from Neolithic sites, and sometimes appeared in a surprising similar form at Bronze Age, Roman Iron Age and Migration Period sites south of the Baltic Sea (Fig. 13), were replaced in Scandinavia and the southern North Sea coast region by three-aisled houses as early as the Middle Bronze Age. The zone of appearance of two-aisled houses is not that well specified and seems to have changed over time. In the west, it is situated to the south of the three-aisled house zone, reaching Westphalia, eastern Brandenburg and parts of Saxony. In Lower Lusatia and south of Berlin, so-called four-aisled houses were discovered (Fig. 14, 63). It is not easy to interpret the plans of these buildings. Here, I present a new proposition for the characteristic post arrangement as supporting a loft (Fig. 64). In the case of one-aisled houses, the inner space is free of posts (Fig. 15) since the walls took over the roof-bearing function. It was a very demanding construction because poor carpentry of joining elements above the wall line inevitably led to its destabilisation and collapse, so it appeared on a larger scale at the beginning of the Middle Ages. However, we also know a few one-aisled longhouses dating to an earlier period. In the next chapter, all elements of the walls are discussed. Special attention is drawn to the fact that rows of posts and walls do not necessarily line up. Since the wall construction is not connected to the house frame or roof, its roof-bearing function can often be excluded (Fig. 20). As the ruins at Feddersen Wierde demonstrate, the line of the wall and that of lateral posts may differ. A special feature are the outer, eave-supporting posts (Fig. 21) that we know from houses in both the west and in the east, but at different times. Such constructions seem to appear in Poland, too. Most of the walls were probably built using the wattle and daub technique. It was predominant used in Central and Northern Europe, but was not the only technique. Houses with wall trenches might have been built with palisade-like walls, with planks (Fig. 26) or as log constructions (Fig. 27). Sometimes there are no traces of the walls at all and the construction must have been over-ground (Fig. 25, 29). With respect to log construction, one drawback is the need for timber, which in regions with limited timber resources can be decisive for choosing another wall variant. For constructing the huge Early Bronze Age house (33.5×ca. 8 m) at Legård on Thy-Vendsyssel (Fig. 27), it was calculated that about 150 oak trees were needed! Most longhouses were built with a rectangular plan, but a quite high number of longhouses in Northern and Central Europe had apse-shaped gable walls (Fig. 30). Roof reconstruction of three-aisled houses with that characteristic seems to pose no problem (Fig. 40–44), but in the case of two-aisled houses with a roof-bearing post in the apse-shaped gable wall, the task of reconstruction is challenging. Regarding the interior structure of Iron Age longhouses, we have a lot of information from the well-preserved house ruins at Feddersen Wierde (Fig. 47–50) and burnt down houses from Denmark (Fig. 51). They prove the widespread use of houses with a living area and stall under one roof. In other cases, the inner division is proven by the existence of small trenches where the partition walls of the boxes were placed (Fig. 52, 53). For now, we cannot determine the precise range of this economic model; the easternmost houses with stall trenches were discovered in Lower Lusatia (right on the German-Polish border). Placing animals under the same roof as people is not a phenomenon limited to antiquity. In some regions of Germany and the Netherlands, it was a fairly common form of farming in modern times. Some of these houses survived until the 1970s (Fig. 54). This type of house was found in a long zone from the vicinity of Amsterdam to the Hel Peninsula – mainly in the zone of the historical range of the Low German language, which is therefore called Niederdeutsches Hallenhaus. At a time when Bronze Age and Iron Age longhouses began to be intensively researched in the Netherlands and Germany, the memory of the original functioning of Niederdeutsches Hallenhaus, so similar to ancient buildings, was still alive, and the grandparents or parents of these researchers often lived in them or knew of such houses anecdotally (Fig. 55:1–3). Some very old buildings showed common structural features with houses from the Roman Iron Age. A comparison of the characteristics of ancient and modern houses has greatly facilitated approaching the subject and interpreting the results of excavations. However, it has sometimes also led to the use of inadequate terms that survive to this day and which are misleading. For example, if the famous researcher of rural architecture J. Schepers talked about Germanisches Hallenhaus or W. Haarnagel in his monumental monograph uses the term dreischiffige Hallenhäuser, they were influenced by the use of almost the same name of the above-mentioned medieval and modern houses that in terms of internal division are so similar to three-aisled longhouses from the Iron Age. However, there is a significant functional difference: the term Halle (hall) in Niederdeutsches Hallenhaus refers to a room with a threshing floor in the central nave, located between livestock bays. This room is large and hall-like, and that is why the houses were given the name Hallenhaus. The ‘hall’ in Late Antiquity (Fig. 58, 59) and medieval times had a completely different meaning and does not mean the same as in the case of rural houses from later times. In the next chapter, I discuss congruencies of house plans as a source of interpretation of incompletely preserved longhouses and for typological divisions. In regard to the latter, we have to take into account the state of preservation, touch-ups, repairs, modifications, extensions and superposition of house plans that influence the interpretation of the record. The same applies to farmsteads and even whole settlements that have been shifted, rebuilt, changed in layout and so on (Fig. 75–80). The issue of forms and structures of settlements is a rather complicated topic, because the condition for their assessment is a completely uncovered site. Such objects are rare, and even if a large complex is excavated, we can only assess the arrangement of objects within the excavations. This statement sounds trivial, but I emphasise this fact because we cannot be sure that there were no satellite units belonging to the given settlement nearby. This is well illustrated by the plan of extremely interesting features at Galsted in southern Jutland (Fig. 81). Its second phase represents another step of settlement evolution and is similar to what we know from settlements such as Nørre Snede in eastern Jutland (Fig. 82). The layout of farmsteads – although already present at some Late Pre-Roman Iron Age sites – represents the state of development of Roman Iron Age and Migration Period settlements. The earliest settlements of this type stem from Jutland, while the tendency to set up large, enclosed rectangular or trapezoidal farms in northern Germany is observable from the late 1st cent. CE and in the northern Netherlands from the 2nd cent. CE. The phenomenon of ‘stationary’ settlements is also known from East Germany, including the already mentioned settlements at Dallgow-Döberitz, Wustermark, Herzsprung or Göritz. Probably such settlements were discovered in Poland, too (see below). Settlements of this type replaced settlements with a different structure, dating to the Pre-Roman Iron Age. Their features included a loose arrangement of farms (rather unfenced) spread out over a large area (Fig. 1) and instability of house and farm sites. Houses and farmsteads were not occupied for a long period of time, but changed relatively quickly (the so-called wandering/shifting settlements). In the Late Pre-Roman Iron Age in Jutland and – in a slightly different form – in the northern Netherlands, completely enclosed settlements appeared. It was a fairly short-lived phenomenon (that ended in the 1st cent. CE), but the first step to stationary settlements, where farmsteads were designed to last for a longer period of time. At sites such as Nørre Snede in Jutland or Flögeln at the North Sea, there was a slow shifting of farmsteads, but over a period of several hundred years. With such a slow pace of changes in the positions of houses and farms, we can actually talk about stationary farms/settlements. It should be emphasised that the structure of settlements during the Roman Iron Age and Migration Period was not compact and there were no clusters of houses around a free square, as is sometimes suggested in Polish literature (admittedly on the basis of insufficient evidence). The image of settlements at that time resembles instead a group of several farms, sometimes in rows. We also know this spatial organisation from settlements in the left-bank regions of the Oder and Neisse Rivers (the German-Polish border) and there is no reason to believe that it was different to the east of these rivers. Despite undeniable progress in recent decades, settlement archaeology in Poland is still at the very beginning of searching for patterns for the recognition and reconstruction of longhouses that can contribute to the determination of individual types. Before completing this stage, analyses at a higher heuristic level do not yet make sense. All attempts to reconstruct settlement structures and search for references in building traditions to other regions in the Barbaricum have ended and often continue to end in failure. There are several reasons for this. First of all, this type of work from the second half of the 20th cent. mainly consisted of incorrect assumptions and axioms – especially regarding the dominance of pit houses in settlements. Secondly, the material that was available cannot create a suitable base for far-reaching conclusions – often the uncovered parts of the settlements were and are still too small to decipher the structures at all; sometimes it is not even possible to say in which part of a given settlement (or farmstead) the researchers conducted excavations. Another, also quite important point is the inaccurate or incompetent recognition of plans for alleged or actually non-existent post houses (Fig. 83). For decades, ‘buildings’ have been published that have no right to exist. Even in contemporary works, we can still find reconstructions (basically recreations) of primitive huts without statics or carpentry rules (Fig. 83), which were exceeded – if they had existed – by longhouses, even in the Neolithic. If buildings were created that have never existed, then obviously the image of a given farmstead must be false, not to mention the settlement structure. The necessity to verify published materials from settlements resulting from the state of research as I have described it does not need to be particularly emphasised. In a sense, the above-mentioned region between the Oder and the Elbe can be a benchmark for Poland. With regard to the state of research on settlements and the research paradigm, the situation in recent decades has been very similar to the situation in recent years in Poland. Until the early 1990s, the regions east of the Elbe could barely contribute to research on the subject of longhouses in the Barbaricum. It seemed that the presence of such buildings at settlements east of these regions that B. Trier (1969) had examined in his basic monograph on Iron Age longhouses was impossible. The very few examples were treated as exceptions. But due to large, often linear investments in infrastructure renewal in the early 1990s, the situation in Eastern Germany changed radically. Suddenly, longhouses started to appear at almost every settlement surveyed. One of the first excavations of this type was carried out in 1994 at the settlement site at Dallgow-Döberitz, a few kilometres west of Berlin, where at least 28 longhouses were discovered, primarily of the three-aisled variety. Publication of research results at Herzsprung in the Uckermark became a milestone, proving in the Oder region the existence not only of three-aisled longhouses, but farmsteads with a layout that was known only until that time from southern Scandinavia and the western part of Central Europe. In 1994–1997, 25 longhouses, mainly two-aisled, were uncovered at Göritz in Lower Lusatia. Today, a similar shift in settlement archaeology is taking place in Poland. Nevertheless, the attempts to distinguish longhouses at settlements in Poland and, at the same time, the frequent lack of experience of archaeologists in this field led to the creation and inclusion of objects that either did not exist in this form or not at all. The biggest obstacle is the lack of models to recognise house types, reflected by the arrangement of posts. There are still very few confidently confirmed three-aisled longhouses in Poland, yet this fact seems to result from the state of research rather than reflect the realities of the Roman Iron Age and Migration Period. To date, we do know four ‘definite’ buildings of this type, three from Pomerania and one from Mazovia; two others houses from central and southern Poland probably also belong to this group: the house I/A at Czarnowo in Western Pomerania (Fig. 85), a not fully uncovered house at Ostrowite in southeastern Pomerania (Fig. 86:1), a house at Leśno in southeastern Pomerania (Fig. 87), and a house in Rawa Mazowiecka (site 38) in western Mazovia (Fig. 88). In my opinion, the traces of a house at Kuców in Central Poland have to be interpreted as two rows of the roof-bearing posts of a three-aisled building (Fig. 89:1), while a house at Domasław in Lower Silesia also probably belongs to the three-aisled type (Fig. 90). Today, we know more examples of two-aisled houses than of three-aisled houses, which primarily appear only in the Przeworsk Culture area. It seems that in fact two-aisled houses were dominant in the area of this cultural unit, but it is still a bit too early to determine this with great certainty. The largest series of longhouses results from excavations of the settlement at Konarzewo near Poznań (Fig. 91), a smaller group we know from the Bzura River region (Fig. 94). The latter form a group that can be used to define the first longhouse type in Poland, the Konotopa type. A very interesting house was discovered in the 1960s at Wólka Łasiecka in Central Poland (Fig. 95). Although the arrangement of the posts is very clear, it can be read in the source publication, and sometimes in later ones, that this building is a three-aisled house. Actually, we are dealing with a two-aisled house with additional, external eave-supporting posts. In the case of the settlement at Izdebno Kościelne in western Mazovia, one can point to a house that was not included in the analysis of the site plan (Fig. 97). The same applies to a two-aisled longhouse at Janków in Central Poland (Fig. 96). It also belongs to the ‘verified’ buildings which were distinguished after the publication of the research results. The above-mentioned house at Wólka Łasiecka can be interpreted as a ‘lime kiln building’ on the basis of similar houses that, for example, were discovered at Klein Köris near Berlin and Herzsprung in the Uckermark. At the latter site, several buildings of this type have been even discovered, at least four of which were longhouses (e.g. Fig. 99:1.6). Lime kiln houses in other forms at this settlement (Fig. 100:3) and subsequent ones (Fig. 99:7, 100:1.2) show that there are many variants of such buildings. It might seem that production halls with limes kilns are a special feature of the settlements of Central Europe from the left-bank regions of the Oder and Neisse to the Vistula. However, the example from Osterrönfeld and houses from the settlement at Galsted in southern Jutland that are not yet published warn against this inference. It is not an exaggeration to claim that previous attempts to distinguish farmsteads in Poland have usually lacked sufficient evidence; often such an activity was and is simply impossible. There are several reasons for this: in the first place, often there are no reliable house plans, also the excavation area is too small and – it should be strongly emphasised – the research results are presented as a schematic plan only or in the form of a plan with symbols. Recently, contrast has been emphasised between the interpretation of the ‘farmstead’ approach among researchers from ‘west of the Oder’ and researchers in Poland, which in my opinion results mainly from the state of research and – probably even in a decisive way – from the research paradigm, and under no circumstances reflects ancient conditions. The results of excavations in recent years have shown that such an contradiction – if used to refer to archaeological material – is only apparent and artificial. The basis for analysing settlement structures in terms of farmsteads is quite narrow, although there are few proposals worth considering. In a separate article, I re-analysed published research results in the area of the settlement at Wytrzyszczki in Central Poland in terms of some longhouses. In addition to the alternative interpretation of buildings, the published plan and field documentation analysis provide the basis for a new interpretation of the spatial organisation of the uncovered part of the settlement (Fig. 102–104). An interesting arrangement of objects was observed at the settlement in at Mąkolice in Central Poland. Both post and pit houses as well as production facilities were uncovered here. The dispersion of all objects is quite clear, but several issues remain an open question (Fig. 105). Closely related to the form of the farmsteads is their arrangement relative to each other, meaning the form of a settlement. Polish literature holds the view that one of the basic forms of settlements of the Przeworsk Culture (because it is the only one we can say anything about) is the circular settlement. The above-mentioned settlement from Wytrzyszczki in Central Poland and well-known settlement from Konarzewo near Poznań cannot be called circular under any circumstances as has happened in the literature (Fig. 104, 106). Concerning the spatial organisation of settlements from areas east of the Oder, I am convinced that they did not differ from settlements in areas west of this river (Fig. 108, 109). The latest field research results provide us with more and more arguments confirming this thesis. The basic unit of each settlement was a farmstead, which was spatially organised as economic units in the western and northern regions of the Barbaricum.