大多数联结主义模型包括 time, i.e. 网络 changes over time。密切联系and extremely common aspect of 联结主义模型 是激活'。任何时候网络的单位 has an 激活, which is a numerical value intended to 描述一些 aspect of 单位。例如，如果模型的单位 are 神经元激活可以描述probability that 神经元 would generate an action potential spike。如果模型 a 传递激活模型then over time a 单位的激活传递给所有与它联结的其他单位。传递激活总是 a feature of 神经网络模型，是认知心理学家经常使用的联结主义模型。
大量研究 utilizing 神经网络is carried out under 更general名称"联结主义"。这些联结主义模型 adhere to 2个主要原则 regarding the mind:
- Any given mental state can be described as a (N)-dimensional vector of numeric activation values over neural units in a network.
- 记忆 is created by modifying 神经单位之间的联结强度。联结强度或 "weights", are generally represented as a (N×N)-dimensional matrix.
尽管有大量variety of 神经网络模型，they very rarely stray from these 2个基本原则。大部分variety comes from:
- Interpretation of units—units can be interpreted as 神经元或神经元groups。
- 激活的Definition—激活 can be defined in a variety of fashions。例如，in a Boltzmann machine，激活 is interpreted as the probability of generating an action potential spike, and it's determined via a logistic function on the sum of the inputs to a unit.
- 学习 algorithm—不同网络 modify their 联结 differently。Generally, any mathematically defined change in 联结 weights over time is referred to as "学习 algorithm".
联结主义are in agreement that recurrent 神经网络(网络 wherein 联结 of 网络 can form a directed cycle)是个更好的大脑模型 than feedforward 神经网络 (网络 with no directed cycles)。大量 recurrent 联结主义模型 incorporate dynamical 系统理论 as well。许多研究者，如联结主义Paul Smolensky, have argued that direction 联结主义模型 will take is towards fully continuous, high-dimensional, non-linear, dynamic systems approaches.
联结主义generally stress 学习的重要性in their models. As a result, many sophisticated 学习 procedures for 神经网络 have been developed by 联结主义。学习 always involves modifying the connection weights. These generally involve 数学公式 to determine the change in weights when given sets of data consisting of activation vectors for some subset of the neural units.
By formalizing 学习 in such a way 联结主义 have many tools at their hands. A very common tactic in 联结主义学习方法 is to incorporate gradient descent over an error surface in a space defined by the weight matrix. All gradient descent 学习 in 联结主义模型involves changing each weight by partial derivative of the error surface with respect to the weight. Backpropagation, first made popular 在1980s, is probably the most commonly known 联结主义 gradient descent algorithm today.
- A set of 处理单位, represented by a set of integers.
- 激活 for 每个单位, represented by a vector of time-dependent functions.
- output function for each unit, represented by a vector of functions on the activations.
- pattern of connectivity among units, represented by a matrix of real numbers indicating 联结强度.
- propagation rule spreading the activations via 联结, represented by a function on the output of the units.
- 激活 rule for combining inputs to a unit to determine its new 激活, represented by a function on the current 激活 and propagation.
- 学习 rule for modifying 联结 based on experience,represented by a change in the weights based on any number of variables.
- environment：provides the system with experience,represented by sets of 激活 vectors for some subset of the units.
大量研究 led to PDP的发展was done in 1970s, but PDP became popular in 1980年代 with the release of 平行分布式处理: Explorations in 认知 Microstructure- 第1册(foundations) & 第2册(心理和生物模型), by James L. McClelland、David E. Rumelhart, 和 PDP研究小组。尽管books are now considered seminal 联结主义 works, the term "联结主义" was not used by the authors to describe their framework at that point。it is now common to fully equate PDP 和联结主义。-->
PDP的直接根源是1950和1960年代弗兰克·罗森布拉特（Frank Rosenblatt）等人的感知机理论。但自从1969年马文·明斯基（Marvin Minsky）和西摩·佩伯特（Seymour Papert）发表了一本名为《感知机》的书以后，感知机（perceptron）模型就变得非常不得人心。明斯基和佩伯特通篇详述了感知机计算功能的局限性。显示甚至无法执行如异或问题（exclusive disjunction，如判断这是苹果还是桔子，但不是二者皆是）这样的简单功能（明斯基后来承认做得过分了）。PDP书克服了这些早期局限性
by showing that multi-level, non-linear 神经网络 were far more robust and could be used for a vast array of 功能。
有许多感知机理论家以外的研究者 who were advocating 联结主义风格的模型 prior to 1980年代。
在1940和1950年代的研究者如Warren McCulloch, Walter Pitts, Donald Hebb和Karl Lashley were advocating 联结主义style理论家。McCullough和Pitts showed how first-order logic could be implemented by 神经系统: their classic paper "A Logical Calculus of Ideas Immanent in神经活动" (1943)很重要in this development here (受到1930年代Nicolas Rashevsky重要工作的影响)。Hebb contributed greatly to speculations about 神经 functioning, and even proposed a 学习原则 that is still in use today, known as Hebbian 学习。Lashley argued for distributed representations as a result of his failure to find anything like a localized engram in years of lesion实验。
联结主义的许多原则可以追溯到心理学的早期工作，如威廉·詹姆士的工作, 尽管 必须指出心理学理论19世纪末对人脑的了解非常时髦。早在1869年，
neurologist John Hughlings Jackson was arguing for multi-level,distributed 系统。Following from this lead, Herbert Spencer的心理学原则, 3rd edition (1872)和西格蒙德·弗洛伊德's Project for a Scientific 心理学(composed 1895) propounded connectionist or proto-connectionist 理论。these tended to be speculative theories。但20世纪初桑代克 was carrying out experiments on 学习 that posited a 联结主义类型网络。
在1950年代的研究者Friedrich Hayek posited the idea of spontaneous order in the brain arising out of decentralized networks of simple units,但Hayek的工作 was generally not cited in PDP literature until recently.
联结主义模型的另一种形式是relational 网络 framework developed by the linguist Sydney Lamb in the 1960s. Relational 网络 have only ever been used by linguists, and have never been unified with the PDP approach. As a result, relational networks are used by very few researchers today。
obliterating the progress made in 认知科学和心理学领域 by 经典计算主义。计算主义是认知主义的特殊形式，argues that 心理活动 is computational, i.e. that the mind is essentially a Turing machine。许多研究者argued that the trend in 联结主义 was towards a reversion to 联想主义, and the abandonment of the idea of a Language of thought, something they felt was mistaken。另一方面，it was those very tendencies that made 联结主义 attractive for other researchers.
联结主义和计算主义 need not be at odds per se, but the debate as it was phrased in the late 1980s and early 1990s certainly led to opposition between the two approaches。throughout the debate some researchers have argued that 联结主义和计算主义 are fully compatible, though nothing like a consensus has ever been reached。二者的不同点 are usually cited are the following:
- 计算主义假定symbolic 模型不resemble underlying 大脑结构 at all, whereas 联结主义 engage in "低水平" modeling, trying to ensure that their models resemble neurological 结构。
- 计算主义 generally focus on the structure of explicit symbols (mental models)和syntactical rules for their internal manipulation, whereas connectionists focus on learning from environmental stimuli and storing this information in a form of connections between neurons.
- 计算主义相信internal mental activity consists of manipulation of explicit symbols, whereas 联结主义相信 the manipulation of explicit symbols is a poor心理活动模型。-->
Though these differences do exist, they may not be necessary. For example, it is well known that connectionist models can actually implement symbol manipulation systems of the kind used in computationalist models. So, the differences might be a matter of the personal choices that 一些联结主义研究者 make as opposed to anything fundamental to 联结主义。
To make matters more complicated, the recent popularity of dynamical systems in 哲学of mind (due to the works of authors 如Tim Van Gelder) have added a new perspective on the debate. Some authors now argue that any split between 联结主义and computationalism is really just a split between computationalism and dynamical systems, suggesting that the original debate was wholly misguided.
All of these opposing views have led to a fair amount of discussion on the issue amongst researchers, and it is likely that the debates will continue.
- 大卫·鲁梅尔哈特（Rumelhart, D.E.）、詹姆斯·麦克莱（J.L. McClelland）和PDP研究小组(1986)：平行分布式处理: Explorations in the Microstructure of Cognition.第1册: Foundations，Cambridge, MA: MIT Press
- 詹姆斯·麦克莱（McClelland, J.L.）、大卫·鲁梅尔哈特（D.E. Rumelhart）和PDP研究小组(1986)：平行分布式处理: Explorations in the Microstructure of Cognition.第2册:心理和生物模型，Cambridge，MA: MIT Press
- Pinker, Steven and Mehler, Jacques (1988). 联结和Symbols, Cambridge MA: MIT Press.
- Jeffrey L. Elman, Elizabeth A. Bates, Mark H.Johnson, Annette Karmiloff-Smith, Domenico Parisi, Kim Plunkett (1996). Rethinking Innateness: A 联结主义 perspective on development，Cambridge MA: MIT Press.
- Marcus, Gary F. (2001). The Algebraic Mind: Integrating 联结主义和认知科学(学习、发展和Conceptual Change), Cambridge, MA: MIT Press
- Dictionary of Philosophy of Mind entry on connectionism
- Stanford Encyclopedia of Philosophy entry on 联结主义
- A demonstration of Interactive Activation and Competition Networks